BACKGROUND: Leisure-time physical activity (LTPA) has been associated with better mental and physical health particularly in cross-sectional studies. Purpose To longitudinally assess the association between LTPA, changes in LTPA and health-related quality of life (HRQL). METHODS: Cross-sectional and prospective analysis of the Seguimiento Universidad de Navarra Project, a dynamic cohort study. Information is gathered through mailed questionnaires biannually since 1999. A validated LTPA questionnaire was used to assess LTPA level at baseline. Changes in LTPA were ascertained at 2 and 4 years of follow-up. HRQL was assessed at 4 and 8 years of follow-up with a validated Spanish version of the SF-36 Health Survey. Multivariate regression coefficients, means and their 95% CIs for each of the eight domains of the SF-36 according to quintiles of baseline LTPA and changes in LTPA over time were calculated using generalised linear models. RESULTS: At 4 years of follow-up, mean scores for upper quintiles of LTPA (second to highest quintile) of physical functioning, general health, vitality, social functioning and mental health were significantly higher than the mean score obtained for participants with the lowest level of LTPA (first quintile). In general, and independent of previous scores in SF-36 survey, the maintenance or the increase in LTPA levels during follow-up was associated with better scores in HRQL after 8 years of follow-up, especially for mental domains. CONCLUSION: These findings provide support for an association between LTPA, long-term changes in LTPA and several aspects of HRQL (especially in relation to mental domains) extending previous cross-sectional findings.