In recent years, accumulating evidence has pointed to the possible negative effects of cannabis use in social and interpersonal context. In this study, we explored the association between level of cannabis use and perceived social support and the extent to which social network size and diversity may mediate this association. Data from the NESARC-III was analyzed. Participants were either non-cannabis users (N = 32,523), using cannabis without qualifying for a diagnosis of cannabis use disorder (N = 2729), or qualifying for Cannabis Use Disorder (CUD; N = 972). Perceived social support was measured by the Interpersonal Support Evaluation List-12 and social network size and diversity were measured using the “Social Network Index.” Baron and Kenny’s statistical method was used for mediation analyses, with control for possible covariates and Bonferroni corrections applied. After controlling for covariates, social network size and diversity independently and fully mediated the association between level of cannabis use and general perceived social support (direct effect: b = −0.02, p = 0.0521, and b = −0.02, p = 0.057, respectively). According to these analyses, level of cannabis use was significantly associated with decreased social network size and diversity, which in turn were associated with lower perceived social support. Our findings highlight the need to address possible changes in social network among cannabis users.