Abstract: Objective To explore the influencing factors of self-injury behavior among undergraduates. Methods A total of 19 662 undergraduates were selected based on stratified random cluster sampling from 10 universities at various levels in Wuhan. Self-injurious behavior questionnaire, emotion adjustment difficulties scale and Eysenck personality questionnaire were conducted among subjects. Binary logistic regression and two-level binary logistic regression model were used to analyze the school-level factors influencing self-injury behavior incidence. Results The incidence of self-injurious behavior among undergraduates was 1.8%. Scores on emotional adjustment difficulties and personality in self-injury group were higher than those in non self-injury group (all P<0.05). The self-injury behavior among undergraduates revealed an aggregation in university-level. The two-level binary logistic regression model showed that gender, origin of student, left-behind experience, emotion dysregulation and personality disorder were predictive factors for self-injurious behavior (all P<0.05). Conclusions It is necessary to build positive coping strategies and improve psychological health level among undergraduates. Peer education intervention plays an important role in the prevention and control of self-injury behavior.