Despite increasing evidence that social capital is positively associated with health, the pathways that link social capital to health are not definitive and invite further investigation. This paper uses household survey data from 22 villages in China in 2002 to test the relationship between social capital and the self-reported health status of the rural population. Focusing on the cognitive dimension of social capital, this paper complements current social capital research by introducing an overlooked distinction between trust and mistrust. Trust and mistrust are measured at the individual and aggregate levels, and the distinct ways in which they affect general and mental health are explored. We adopt an ordered logistic regression using survey procedures in SAS version 9.1 to account for the stratified and clustered data structure. The results suggest that: (1) individual-level trust and mistrust are both associated with self-reported health in rural China--trust is positively associated with both general health and mental health, while mistrust is more powerfully associated with worse mental health; and (2) the effects of individual-level trust and mistrust are dependent on village context--village-level trust substitutes for individual-level trust, while individual-level mistrust interacts positively with village-level mistrust to affect health. However, an unexpected protective health effect of mistrust is found in certain types of villages, and this unique result has yet to be examined. Overall, this study suggests the conceptual difference between trust and mistrust and the differential mechanisms by which trust and mistrust affect health in rural China. It also suggests that effective policies should aim at enhancing trust collectively or reducing mistrust at the personal level to improve health status in rural areas of China.