OBJECTIVES: To investigate direct and stress-buffering associations between social support from family and the mental health of older people in Iran, a country which has recently undergone an exceptionally fast fertility transition and is consequently experiencing rapid population ageing. METHOD: A cross-sectional stratified random survey of 800 people aged 60+ years resident in Tehran was conducted. In total, 644 people responded. The Social Provisions Scale and the General Health Questionnaire were used to measure perceived social support and mental health, respectively. Multilevel mixed-effects models were used to examine the hypotheses. RESULTS: The findings supported the hypothesis of a direct association between perceived and received social support and mental health. However, we did not find strong evidence to suggest that social support buffered the effects of stress arising from limitations of physical functioning. Lack of help doing paperwork was associated with worse mental health for women but not men. Source of support did not seem to be important. CONCLUSION: Our results indicated that in Tehran, as in Western settings, social support is important for the mental well-being of older people. Recommendations for policy and further research priorities based on the study findings were provided.