Australian mental health policy emphasises the central importance of social inclusion for promoting mental health and treating illness. Social capital is a commonly used concept for describing and evaluating social inclusion, but its relationship to mental health has not been investigated among groups of differing cultural backgrounds in rural and remote Australia, where need is greatest. The aims of this study were to compare levels of social capital and associations with mental health among Aboriginal Australians, New Australians and Other Australians living in a disadvantaged coastal community and to draw implications for policy and service development. Participants were 963 community members, aged 19-97, randomly selected from a socio-economically disadvantaged coastal Australian region, voluntarily completing an anonymous postal survey. Measures of components of social capital were analysed in terms of general psychological distress and happy feelings using multiple hierarchical linear regression modelling. For all groups, higher levels of participation, more positive perceptions about participation, and greater cohesion were associated with less distress and more happy feelings. In multivariate analyses, greater social capital overall was strongly associated with less distress and more happy feelings, the former particularly for Aboriginal Australians. Different aspects of social capital were related to mental health and wellbeing for different groups. Social capital is essential for good mental health and wellbeing but it is neither evenly shared among different groups nor do the same components of social capital matter, or matter to the same degree, among different groups. Social capital may be particularly important (and sensitive) for protecting Aboriginal people against mental health problems. In that these findings suggest approaches to designing policy and services, these must be tested via careful intervention research.