At present, more and more adolescents use drug. More than 30% reform school students use drug. Drug use greatly endangers adolescents’ health. Therefore, we should pay more attention to drug use among reform school students. In order to design effective prevention and intervention programs, we should study the potential mechanisms of drug use of reform school students. Due to the signi?cant biological, psychological, and social changes during adolescence, individuals have to face many challenges from family, school, and peer settings. Many studies have suggested that stressful life events play an important role in the development of adolescents’ drug use (Andersen & Teicher, 2009; Kulis, Marsiglia, & Nieri, 2009; Ye, Wen, Yang, Chen，Zhang, in press). Although stressful life events have been found to account for large proportions of variance in reform school students’ drug use, they do not lead to the same extent of drug use for all reform school students. The impact of stressful life events may be contingent upon many factors. Unfortunately, earlier studies on the effects of stressful life events have typically investigated direct relationships between the independent variables (i.e., stressful life events) and dependent variables (i.e., drug use) without exploring the processes that may underlie these relationships. The present study ?lled this gap by incorporating family functioning into the examination of the effects of stressful life events on reform school students’ drug use. In other words, family functioning may moderate the relationship between stressful life events and drug use. Although the interactions between stressful life events and family functioning may plausibly explain the differences of reform school students’ involvement in drug use, the question remains how family functioning may in?uence the relations between stressful life events and drug use, or, in other words, what process is responsible for the moderation effect. In the present study, we proposed a mediated moderation model to examine potential underlying mechanisms of the stressful life events ×family functioning interactions predicting drug use. In a mediated moderation model, a moderating effect can be explained, at least in part, by a mediator variable. More precisely, it was hypothesized that perceived social support mediated the moderating effect of family functioning on the relationship between stressful life events and drug use. The present study aimed to investigate whether family functioning moderated the relationship between stressful life events and drug use, and perceived social support mediated this moderation effect after controlling for several demographic variables. A sample of 514 reform school students of 10 reform schools (427 boys and 87 girls, Mage=14.25±0.92) was recruited in the study to complete self-report questionnaires. The self-report questionnaires used in this study included stressful life events scale, family function scale, perceived social support questionnaire, and drug use questionnaire． Results were as follows: (1) family functioning moderated the relationship between stressful life events and drug use. (2) perceived social support mediated the moderation effect. Perceived social support could partially mediate the relationship between stressful life events × family functioning and drug use. The effect of stressful life events on drug use among reform school students is mediated moderating effect.