Tracking and explaining physical activity in young adults was investigated in a 7-year longitudinal design. From a representative sample of 980 respondents at baseline (M age = 21 years), 172 respondents, comprising 18% of the original sample, completed psychosocial questionnaires and were interviewed at home to assess their level of physical activity at follow-up. Psychosocial variables were grouped into four groups: social variables, self-efficacy, perceived benefits, and perceived barriers. Drop-out analyses showed no major differences between the baseline and follow-up samples. The aims of the present study were: (a) to investigate tracking in physical activity and psychosocial determinants over a 7-year period, and (b) to examine baseline determinants and change in determinants in predicting physical activity change in men and women. No significant tracking of physical activity level was found for men. For women, moderate tracking scores were found for total energy expenditure and moderate-intensity energy expenditure (r = .34 and .41, respectively). There was higher tracking in psychosocial determinants than in physical activity. Relatively high tracking scores for perceived benefits and barriers suggest that cognitions related to physical activity are more stable than the behavior itself. Although baseline psychosocial variables were poor predictors of physical activity change, determinants' change scores accounted for 16-19% of the variance in physical activity in men and 7-24% in women. Significant predictors were different for men and women. Present results can inform the design of physical activity interventions for these populations.