Abstract Background Although social support has been observed to exert a beneficial influence on leisure-time physical activity (LTPA), multidimensional approaches examining social support and prospective evidence of its importance are scarce. The purpose of this study was to investigate how four dimensions of social support affect LTPA engagement, maintenance, type, and time spent by adults during a two-year follow-up. Methods This paper reports on a longitudinal study of 3,253 non-faculty public employees at a university in Rio de Janeiro (the Pró-Saúde study). LTPA was evaluated using a dichotomous question with a two-week reference period, and further questions concerning LTPA type (individual or group) and time spent on the activity. Social support was measured by the Medical Outcomes Study Social Support Scale (MOS-SSS). To assess the association between social support and LTPA, two different statistical models were used: binary and multinomial logistic regression models for dichotomous and polytomous outcomes, respectively. Models were adjusted separately for those who began LTPA in the middle of the follow up (engagement group) and for those who had maintained LTPA since the beginning of the follow up (maintenance group). Results After adjusting for confounders, statistically significant associations (p