With the increase in the proportion of older adults in our country, the psychological problems of older adults have aroused the attention from academia and the whole society. Maintaining a high level of mental health is directly related to the overall welfare levels and living conditions of older adults. The scholars abroad have begun to explore the relationship between social participation, especially participation in commonweal activities, and mental health of older adults. However, in China participation in commonweal activities of older adults have not yet formed a broad and effective pattern. In this context, we attempt to study the impact of participation in commonweal activities of older adults on mental health, and hope to provide new ideas for the research on alleviating the pressure of aging population and achieving the happiness of older adults. By reviewing the related literature about participation in commonweal activities and mental health of older adults, we forwarded the theoretical model in which participation in commonweal activities could affect mental health, and hypothesized that participation in commonweal activities can highly improve the mental health of older adults and provide ″social compensation effect″ for vulnerable people. If vulnerable older adults participate in commonweal activities, they can get more benefit. On this basis, by using the data from Chinese General Social Survey (CGSS) in 2005, we made an empirical analysis. First, we found that education, income, health and being a communist had a positive effect on participating in commonweal activities among older adults. Older adults with higher socioeconomic status were more likely to participate. Then we used the treatment-effect model which could control and eliminate sample selection bias to find how participation in commonweal activities influences mental health. We found participation in commonweal activities can highly improve the mental health of older adults. Older adults who participated in commonweal activities reported higher level of mental health. Finally, we chose health, income and education to measure the vulnerability of older adults, and used a treatment-effect model with interaction term to prove ″social compensation effect.″ The findings show that when older adults with poorer health, lower education and lower income participated in commonweal activities, their level of mental health could get more improvement. In other words, vulnerable older adults could get more benefit from participating in commonweal activities, which strongly prove ″social compensation effect.″ We combined participation in commonweal activities and mental health, which provided a useful supplement for the current research field of older adults. And we creatively found that participation in commonweal activities could provide ″social compensation effect″ for vulnerable people. Although there are challenges to recruiting vulnerable older adults to participate, their psychological problems should gain more attention from the whole society. Ignoring them would produce the Matthew effect. Thus, we suggest the whole society may encourage older adults, especially vulnerable older adults, to participate in commonweal activities.