The present research aimed at predicting students' social capital based on contextual variables. Theoretical framework chosen for this study was from the works of Putnam, Coleman, Bourdieu, Fild and Fukuyama. The method used was correlational type. The subjects included 400 students of Islamic Azad University, Zone 11, selected by multi-level sampling. Tools of the research were the Social Capital Questionnaire, and researcher made questionnaire. The results of data regression analyses indicated that many of contextual variables can predict social capital factors (i.e., general confidence, formal-civil participation, and religious participation, norm of bilateral dealing, friendly-neighborhood relationship, friendly-family relationship, and organizational confidence). The results showed that: organizational confidence in women is more than men; job can only predict formalcivil participation; formal-civil participation is higher among graduate students; field of study can generally predict social capital; significant differences were among all variables, except organizational confidence, and; general confidence and formal-civil participation were more in married students than single students.