Jeju Woman Killed After Four Reports to Police, With Safety Watch Torn From Her Wrist

Four reports preceded the fatal attack

A woman in her 40s who had reported her former partner to police four times over stalking, threats involving intimate footage and dating violence was fatally stabbed in Jeju City after her police issued emergency smartwatch was torn from her wrist. Investigators found the damaged device near the entrance to her home.

Jeju Seobu Police Station said the suspect, a man in his 60s, waited for the woman to return home before grabbing her and dragging her toward his vehicle. Most accounts place the attack at 1:38 a.m. on October 12, outside a hotel in Yeon-dong. Police arrested him at 5:08 a.m. in Geonip-dong, three hours and 30 minutes later.

The four reports covered two stalking incidents in October of the previous year, a complaint involving sexual violence on March 5, and dating violence in May. None resulted in an order restricting the man's access to the woman. Police nevertheless assessed her risk, monitored her periodically and provided a smartwatch and personal alarm in September.

The case exposes a distinction that matters to understanding her protection: police had registered the woman for safety assistance, but the earlier complaints had not led to continuing criminal proceedings or a restraining order. Investigators are now examining the killing, while the handling of those complaints has drawn scrutiny.

What happened to the earlier complaints?

According to the more detailed police account, the two October stalking reports ended with action at the scene and warnings after the woman said the situation had been resolved and did not want a formal case. The May dating violence complaint concerned the man visiting her workplace or interfering with her work.

Police said officers interviewed her in May and encouraged her to proceed with a formal complaint and protective measures, but she declined further action. That account does not explain why she declined or establish that the danger had ended.

The March 5 complaint followed a different course. It was formally registered and concerned an allegation that the man had threatened to distribute a video of sexual intercourse. Police said the woman later submitted a statement saying she did not want him punished and did not attend questioning. Investigators closed the case for insufficient evidence.

This detail qualifies the initial description that all four matters ended at the scene, and other accounts suggesting none became formal cases. The March complaint did enter an investigation before being closed. Accounts also differ on the number of May reports: several mention multiple contacts that month, while the detailed police tally counts one May complaint within a total of four. The distinction between a call for help and a formally registered criminal case is central to the police explanation.

A timeline of reports and protection

The sequence stretches across roughly a year, with further unwanted contact reported after the earlier complaints had ended. One account explicitly dates the March complaint to 2026, while another dates the preceding October stalking reports to 2025. On that chronology, the September protection measures and October killing occurred in 2026.

• October 2025: The woman reported stalking twice. Police said the incidents ended with warnings or action at the scene rather than formal criminal cases.
• March 5, 2026: Police registered the complaint involving threats to distribute intimate footage. It was later closed for insufficient evidence.
• May: She reported dating violence, including unwanted workplace visits or interference with her work.
• September 15: Police reportedly conducted an interview during monitoring.
• September 16: Police issued a smartwatch and registered her for safety management in the 112 emergency reporting system.
• September 22: Police provided a personal alarm.
• September 28: An account of workplace monitoring says an alarm sounded when the man appeared and he fled.
• October: The man reportedly made three unwanted telephone calls to her.
• October 12: Most accounts place the fatal attack at 1:38 a.m. and the arrest at 5:08 a.m.

Reports differ on the date of the killing. One English account places it on October 13, while the detailed Korean accounts and several other reports give October 12. The October 12 sequence is also tied to the reported arrest time. That discrepancy should not be treated as a second incident.

How the smartwatch was removed

Early accounts said the woman was wearing the smartwatch but could not activate it during the sudden assault. Later details indicate that it was pulled off near her home before she was taken toward the site of the stabbing.

The suspect reportedly told investigators that he pulled at her arm when she touched the wrist carrying the device. Police found the watch with its fastening damaged near the building entrance. One account says she was dragged about 100 metres from her residence toward the roadside attack location.

The initial account describes the suspect as admitting that he removed the watch to prevent a report. A more detailed police account is less definite: investigators said they had not yet established whether he knew in advance that she wore the device or understood its emergency function. His reported explanation and the police assessment should therefore be distinguished.

The smartwatch was designed to send an emergency report to the 112 control room when its button was pressed and allow police to identify the wearer's location. No emergency report was made through it during the attack. The evidence described so far concerns its removal and the woman's inability to use it, not a confirmed electronic malfunction.

An attack on a busy street

Investigators believe the man waited near the woman's home for her return, grabbed her and forced her toward his parked vehicle. Accounts describe the stabbing as occurring outside a hotel in Yeon-dong, close to her residence. They do not establish every step of the movement between the building, vehicle and roadside.

Witnesses who called 112 reportedly described a man dragging a woman along the pavement while assaulting her. They also said he left her lying beside the road and drove away. She was taken to hospital but died from her injuries.

The location was beside a road with four traffic lanes, near convenience stores, restaurants and nightlife businesses. Reports describe an area where people and taxis commonly remain active late at night. Witness calls brought police to the scene, but they did not prevent the fatal injuries.

The suspect was arrested in Geonip-dong after fleeing. One account places the arrest location about six kilometres from the attack. Police are examining surveillance recordings, vehicle camera footage and his movements to establish the sequence more precisely.

What the risk assessment provided

Police classified the woman as a B grade protection subject, with monitoring once every two months. A report describing the grading system says B grade applies where emergency reports and registered cases each occur at least twice in a year. It describes A grade as involving at least three of each and monitoring once a month.

Those reported criteria do not, by themselves, explain this woman's classification. Several of her complaints did not become formal cases, and the full assessment record has not been disclosed. It is therefore not possible to reconstruct the decision simply by counting the four reports.

The September measures show that police continued to recognise a safety concern even after the complaints had ended. Alongside issuing the smartwatch and personal alarm, they registered her in the 112 system because of concern about retaliatory violence.

The reported September 28 workplace incident is particularly relevant. An account says the personal alarm sounded when the man appeared and he fled. That suggests the alarm interrupted one encounter, but it did not prevent later contact. The record does not establish whether this incident changed her risk grade, monitoring schedule or the consideration of restrictions on the man.

Why the absence of a restraining order matters

Police said the closure of the earlier complaints, the woman's reluctance to pursue further action and the lack of sufficient statements or other evidence limited their ability to seek restrictions. No order barring the man from approaching her was issued.

Several accounts also point to a change in South Korea's Act on Punishment of Stalking Crimes: the provision preventing punishment against a victim's expressed wishes had been abolished. They describe the revised law as requiring police to handle a stalking case even when the victim does not want the alleged offender punished.

A request not to punish someone and a shortage of evidence are different issues. Removing the victim's ability to block prosecution does not remove the need to establish an offence. It does, however, mean that unwillingness to seek punishment cannot alone explain why a stalking allegation was not pursued.

The unresolved issue is what police assessed and documented at each stage. The fact that the woman declined some measures does not establish that all possible investigative or protective options had been exhausted. Equally, the reported change in the law does not prove, without the case records, that a particular order would necessarily have been granted.

The contrast in this case is concrete: the woman had equipment that could summon help during an emergency, but no reported order restricted the man's approach beforehand. Those are different forms of protection, and the equipment depended on her being able to use it.

What investigators are examining next

Police are investigating both motive and whether the killing was planned. The suspect reportedly said he acted after the woman refused to meet him. Investigators are also examining his claim that he had lent money to one of her acquaintances through her and wanted repayment.

The existence and terms of that alleged debt have not been established in the reported investigation. His account is a claim about motive, not a confirmed explanation of the events or any justification for the attack.

Police are also investigating reports that he bought the weapon about a month before the killing. A purchase before the attack may be relevant to planning, but investigators had not announced a conclusion on that point.

The announced work includes examination of mobile phone data, surveillance footage, vehicle recordings and travel routes. Police said they planned to seek an arrest warrant on suspicion of murder and conduct an autopsy on October 13. No warrant decision, autopsy result, trial date or completed finding on planning is established in these accounts.

National figures show repeated requests for help

The Jeju killing comes alongside figures from the National Police Agency submitted to Seo Myeong-ok, a People Power Party member of the National Assembly's Public Administration and Security Committee. Stalking reports rose from 14,509 in 2021 to 44,687 in the most recent completed year cited, more than tripling.

Dating violence reports increased from 57,305 to 105,327 over the same period, a rise of 83.8%. The wider category covering stalking, dating violence, domestic violence and child abuse rose from 316,542 to 475,639 reports, an increase of 50.3%. A further 358,768 reports were provisionally recorded from January through August of the reporting year.

The repeat reporting figures are especially relevant, but require care. Of 19,316 repeat stalking reports, 15,090, or 78.1%, were in the category involving at least three reports. For dating violence, the corresponding figures were 25,322 out of 32,634, or 77.6%. Together, 40,412 of 51,950 repeat reports fell into that category, about 77.8%.

These statistics are grouped by victim and can include different alleged offenders or an unidentified offender. They do not show that every repeat report concerned the same person. They demonstrate repeated requests for help, rather than a precise count of repeat offending by individual suspects.

Police smartwatch provision rose from 10,989 in 2021 to 21,974 in the latest completed year, almost doubling. Yet the reported staffing figures list 259 officers dedicated to victim protection across 262 police stations. Stations in the third administrative grade reportedly had no dedicated officer, leaving other personnel to combine protection with existing duties.

Seo called for police to intervene earlier when danger signs appear and for stronger separation and measures to prevent further offences by suspects assessed as posing a high risk. The national figures support scrutiny of capacity and repeat reporting, but they do not establish whether staffing caused any specific decision in Jeju. That requires the local records.

Key Points

• The woman had made four reports involving stalking, threats concerning intimate footage and dating violence before the killing.
• The March complaint became a formal investigation but was closed for insufficient evidence. The other complaints ended without continuing criminal cases.
• Police issued a smartwatch on September 16 and a personal alarm on September 22, but no restraining order was imposed.
• The smartwatch was found torn off near her home. Investigators had not confirmed whether the suspect knew its function in advance.
• Most accounts place the attack at 1:38 a.m. on October 12 and the arrest at 5:08 a.m.; one account gives October 13 for the killing.
• Police were investigating motive, an alleged debt, the weapon purchase and possible planning, and intended to seek a murder arrest warrant.
• National repeat reporting figures show frequent requests for help, but do not always concern the same alleged offender.