Previous studies have shown that participation in leisure time physical activity is related to better mental well-being and subjective health. However, the associations between different types of leisure time physical activities and different dimensions of mental well-being have rarely been studied. In addition, longitudinal research, analyzing possible causal relations between these variables, is lacking. To investigate these research questions, data gathered at ages 42 and 50 (present N = 303) for the Finnish Jyväskylä Longitudinal Study of Personality and Social Development were used. Physical activity was assessed as frequency of participation at ages 42 and 50, and at age 50 also as frequency of participation in different types of physical activities. Mental well-being was captured by emotional, psychological and social well-being and subjective health by self-rated health and symptoms. Cross-sectionally, different types of physical activities were related to different dimensions of well-being. Walking had positive associations with psychological and social well-being, rambling in nature with emotional and social well-being, and endurance training with subjective health. Rambling in nature was also positively related to subjective health but only among men. Longitudinally, mental well-being predicted later participation in leisure-time physical activity, whereas no longitudinal associations between subjective health and physical activity were found. The results suggest that leisure time physical activities are related to current mental well-being and subjective health in midlife. Across time, good mental well-being seems to be a resource promoting engagement in physical activity.