Previous studies show migrants are generally healthier than the populations in receiving societies, a result generally attributed to the positive selection of migrants on health. This hypothesis, however, has not been adequately evaluated due to lack of adequate data. In this article, using high-quality longitudinal data from Indonesia, the health selectivity hypothesis, also referred to as the healthy migrant hypothesis, is examined with respect to internal migration. Specifically, this study explores whether pre-migration health status affects the likelihood of migration by comparing those from the sending population who do and do not move. Results show that migrants in Indonesia tend to be selected with respect to health and that this selection is robust to household unobserved heterogeneity. However, the strength and direction of the health-migration association vary by types of migration and dimensions of health.