This secondary analysis of the Health Survey for England 2000 aimed to investigate whether individual-level social capital is associated with care home residence and with function, mental health, and self-assessed health in older adults. Older adults in both care home and community residential settings were included. Two indicators of social capital, perceived social support and group participation, were considered for both care home and community-dwelling respondents. Amongst community dwellers, trust in others was considered as a third indicator. Functional impairment, psychiatric morbidity, and self-assessed health were used as indicators of health. Multivariable modelling was undertaken using logistic or ordinal logistic regression. The results show that severe lack of social support was associated with over twice the odds of care home residence, with increased odds of psychiatric morbidity in both care home and community settings, and with more severe functional impairment and worse self-assessed health in the community but not in care homes. Participation in more groups was associated with lower odds of functional impairment in both settings, and with lower odds of psychiatric morbidity and better self-assessed health among community but not among care home respondents. High levels of trust were associated with lower severity of functional impairment, reduced odds of psychiatric morbidity, and better self-assessed health. It is concluded that individual-level social capital was associated with care home residence and with indicators of physical, mental and self-assessed health. These associations differed between community and care home settings, and were generally stronger in the community.