OBJECTIVE： To provide reference for developing effective interventions to self-medication and promoting reasonable self-medication. METHODS： Questionnaires were designed based on the modified Anderson’s behavioral model of health services. By using cluster sampling， residents were randomly selected from 4 districts of Wuhan city to analyze the self-medication behavior and its influential factors. RESULTS： Totally 204 questionnaires were sent and 199 were recovered with effective rate of 97.5％. The result showed 47.7％ residents who perceived discomfort within 2 weeks medicated themselves， and the rate was higher than “see the doctor”. Among self-medication residents， the rate of mild disease， general and serious disease was separately 50.5％， 46.3％ and 3.2％. The length of illness＜7 d and the length of illness ≥7 d  accounted for 95.8％ and 4.2％， separately. Cold and cough was the highest cost disease for self-medication. The reasons for selecting self-medication were： no need to “see the doctor” （43.2％）， trouble-saving （23.1％）， high medical costs （14.6％）， no time to see the doctor （14.6％）. 54.1％ residents acquired basic medication knowledge from the past experience， 25.5％ from other’s recommendation， and the other from the internet， newspaper and magazines. As for the effect of self-medication in the past one year， 75.4％ residents considered most were effective. 18.6％ had suffered from ADR due to self-medication； the incidence of ADR in the residents reading the instruction was significantly lower than those not reading the instruction（P＜0.05）； the ratio of reading the instruction in residents with low education level was significantly lower than that with high education level（P＜0.05）. Single factor analysis showed that residents’ choice of self-medication were significantly influenced by the marital status， education， monthly income， medicare， accessibility of medical institutions， the severity of disease and duration of illness（P＜0.05）.  The proportion of self-medication for married group was significantly higher than unmarried group； the proportion of self-medication for low education level group was significantly higher than high education level group； the proportion of self-medication for low monthly income group was significantly higher than high monthly income group； the proportion of self-medication for group with business medical insurance or self payment was significantly higher than group with basic medical insurance and public medical insurance； the proportion of self-medication for group far from medical institutions was significantly higher than that close to medical institutions； the proportion of self-medication for mild disease group was significantly higher than serious disease group； the proportion of self-medication for length of disease≥7 d was significantly lower than length of disease＜7 d. Results of multivariate analysis showed that severity of disease and length of disease were the important influential factors for the choice of self-medication（P＜0.05）.  CONCLUSIONS： Self-medication is common in Wuhan city， and mainly for general and chronic disease. Residents don’t pay enough attention to reading the drug instructions. Self-medication is affected by various factors which has a certain health risks.