Objective: To identify the latent classes of various health risk behaviors among medical students and further analyze the correlation between different classes and depression. Methods: Using the method of convenient sampling, 2 014 medical students were selected from Anhui Medical University in October 2018. Self-designed online questionnaire were used to collect demographic data, health risk behavior (including smoking, drinking, suicide, sleep disorders, mobile phone dependence and lack of exercise) and depressive symptoms. Latent Class Analysis was used to cluster individuals. Logistic regression was provided to further analyze the association between health risk behaviors and depressive symptoms for the groups. Results: The health hazard behaviors of medical students could be divided into three separate classes: substance dependence group (8.4%), mobile phone dependence group (22.4%) and low-risk group (69.2%). The distribution of health risk behaviors of medical students with different gender, major, grade, only child，father s educational level，monthly living expenses, academic achievement and number of friends were statistically significant (χ 2=99.37,19.07,12.05,6.64,14.28,19.35,20.61,26.39,P<0.05). The results of Logistic regression analysis showed that after adjusting for gender, major, grade, only child, father s educational level，monthly living expenses, academic achievement and number of friends, the mobile phone dependence group was positively correlated with depressive symptoms (β=1.75, P<0.01). Conclusion Different ratent classes of health risk behaviors have different correlation with depressive symptoms in groups.It is suggested that school health workers should make individualized intervention plan for different types of health hazard behavior of medical students in the future educational activities, carry out stratified intervention, alleviate the symptoms of depression and promote the mental health of medical students.