OBJECTIVE: to identify socio-psychological predictors of mortality during a 20-year follow-up period among people aged 65 to <85 and 85+ at baseline interview. STUDY DESIGN AND SETTING: elderly people living at home in East London and mid-Essex, who responded to surveys of successful survival in older age in the late 1980s; their mortality was traced through the National Health Central Registry. RESULTS: adjusted analyses show that, as expected, the hazard rate for mortality over a 20-year follow-up was reduced for younger respondents and increased for less functionally able respondents. The hazard ratio for males was almost one and a half times that of females. The hazard rate was also reduced with each categorical increase in life satisfaction and was consistently reduced for those who undertook crafts, social visiting and activities regularly. There was some variation by age and sex. CONCLUSION: the results show that social participation is associated with lower risks of death, particularly among people aged 65 to <85, and that life satisfaction is also protective, particularly among females and people aged 85+, even when health status and socio-demographic circumstances are controlled. The study thus provides support for the hypothesised influence of social participation and subjective well-being on survival in older age.