Abstract Background Although suicide has been postulated as a result of social breakdown, relatively little attention has been paid to the association between social relationships and non-fatal self-harm. We sought to investigate the extent to which social factors correlate with self-harm in this case–control study. Methods The primary outcome was self-harm with hospital presentation. Cases of self-harm from the Emergency Department in a general hospital in Northern Taiwan were recruited, and individually age-and-gender-matched control participants were recruited from non-psychiatric outpatient clinics at the same hospital. The Close Persons Questionnaire was administered and its social support and social network subscales were used to measure social relationships in the 12 months prior to the interview. Other covariates, comprising sociodemographic factors, major life events, physical and mental health, were adjusted in conditional logistic regression models. Results A total of 124 case–control pairs were recruited. The mean (standard deviation) age of the case group was 34.7 (12.8) years and 80.6% were female. Higher social isolation score remained significantly associated with self-harm after adjustment (adjusted odds ratio per standard deviation increase 2.92, 95% confidence interval 1.44-5.95) and household size was negatively associated with the outcome (adjusted odds ratio per unit increase 0.54, 95% CI 0.32-0.94). Conclusions More limited social networks were associated with self-harm after adjustment for potential confounders. Enhancing social structure and effective networking of people with self-harm to community resources may be important for self-harm management in Asian societies and elsewhere.