OBJECTIVES: Successful aging (SA) refers to maintaining wellbeing in old age. Several definitions or models of SA exist (biomedical, psychosocial and mixed). We examined the longitudinal association between various SA models and socio-demographic factors, and analyzed the patterns of change within these models. METHODS: This was a nationally-representative follow-up in Spain including 3,625 individuals aged ≥50 years. Some 1,970 were interviewed after three years. Linear regression models were used to analyze the survey data. RESULTS: Age, sex and occupation predicted SA in the biomedical model, while marital status, educational level and urbanicity predicted SA in the psychosocial model. The remaining models included different sets of these predictors as significant. In the psychosocial model, individuals tended to improve over time but this was not the case in the biomedical model. CONCLUSIONS: The biomedical and psychosocial components of SA need to be addressed specifically to achieve the best aging trajectories.