Depression is associated with social dysfunction and maladaptive social environments, but mechanisms through which social relationships affect depressive psychopathology are unclear. We hypothesized that emotion regulation (ER) is such a mechanism, with outcomes of individuals’ ER efforts sensitive to the social context, and individuals’ ER strategy repertoire and use sensitive to social influence. In Study 1, a longitudinal study of community adults (N = 1,319), associations of individuals’ ER strategies with depressive symptoms depended on social connectedness and romantic relationship status (social context hypothesis). Moreover, associations of social connectedness and relationship status with symptoms were accounted for by maladaptive ER concurrently and, for social connectedness, prospectively over 1 year (social influence hypothesis). Study 2a, using a national sample (N = 772), replicated and extended these findings with a broader array of ER strategies, and ruled out alternative explanations regarding social skills and psychological wellbeing. Among participants in romantic relationships (Study 2b; N = 558), intimacy and trust buffered associations of maladaptive ER strategies with symptoms (context), and maladaptive and adaptive ER mediated links between relationship variables and symptoms (influence). Findings suggest that close relationships—and variation in underlying relational processes within relationships— influence the ER strategies people use, and also affect whether individuals’ own ER repertoires contribute to depression when deployed. Results elucidate core social mechanisms of ER in terms of both basic processes and depressive psychopathology, suggest ER is a channel through which social factors affect internal functioning and mental health, and inform relationship pathways for clinical intervention.