Background and Objective: The aim of this study was to evaluate psychosocial predictors of substance use during adolescence among the Iranian youth. Based on the “peer cluster theory” (Oetting and Beauvais 1986a, 1986b), structural connections between family, school, peer cluster, and religion, as well as their direct and indirect effects on drug use were studied. Materials and Methods: In this study, we collected data on 746 school students from 19 different schools in 9 educational districts of Tehran. Using confirmatory factor analysis, the validity of scales were evaluated, followed with structural equation modeling analysis using LISREL (8.53 package), generalizability of “Peer cluster theory” evaluated the findings. Results: The results showed that our scales were reliable in measuring latent variables in the model. Based on fitness indices, Peer cluster theories with some modification in endogenous variables have good generalizability among the Iranian adolescents. Conclusion: We can consider the "Family" as a fundamental factor in prevention of drug use in adolescents. As long as the family has less power, e.g. paying less attention to their adolescent or the presence of conflicts between parents, the possibility of drug use in adolescent will increase.