To explore the detected characteristics of non-suicidal self-injury (NSSI) behaviors and their relationship with auditory hallucination in patients with depressive episodes. Methods: A total of 473 patients with depressive episodes were enrolled by random cluster sampling. The general demographic questionnaire, the Ottawa Self-Injury Scale and the Chinese version of the Psychiatric Symptom Rating Scale (PSYRATS) were used to analyze the relationship between NSSI and auditory hallucination symptoms. Results: The detection rate of NSSI in patients with depressive episodes was 40.6% . NSSI had statistically significant differences in age, sex, course of disease, occupation, education level, marital status, family history, diagnostic classification, smoking and drinking history, and presence of auditory hallucinations (P<0.05). The prevalence of NSSI was 42.9% and 67.6% in men and women with auditory hallucinations, and 19.8% and 42.1% in men and women without auditory hallucinations; the differences were statistically significant (P<0.05). In terms of diagnostic classification, the detection rate of NSSI between patients with and without auditory hallucinations in the unipolar depressive episode group was significantly different (P<0.05), which was significant different both in males and females (P<0.05). The detection rate of NSSI between patients with and with auditory hallucination in the group of bipolar depressive episodes was significantly different (P<0.05), which was significant different only in females (P<0.05). Auditory hallucinations were associated with the most frequent sites of NSSI (P<0.05). After adjustment of for gender, age and occupation, binary Logistic regression showed that there was a significant positive correlation between auditory hallucinations and NSSI behavior characteristic values (OR=1.04, 95%CI 1.01～1.08, P<0.05). Conclusion: Auditory hallucinations in patients with depressive episodes is positively correlated with NSSI behavior.