The purpose of this study is to examine the reciprocal relationship between motor abilities and physical activity and the mediation effects of physical self-concept in this relationship using longitudinal data. We expect that the effects of motor abilities on physical activity are rather indirect via physical self-concept and that the effects of physical activity on motor abilities are rather direct without involvement of the motor ability self-concept. Data was obtained from the Motorik-Modul (MoMo) Longitudinal Study in which 335 boys and 363 girls aged 11-17 years old at Baseline were examined twice in a period of six years. Physical activity was assessed by the MoMo Physical Activity Questionnaire for adolescents, physical self-concept by Physical Self-Description Questionnaire and motor abilities by MoMo Motor Test which comprised of the dimensions strength, endurance, coordination and flexibility. Multiple regression analyses were used to analyse the direct and indirect effects. The results of the multiple regression analyses show that the effects of motor abilities on physical activity were only indirect for the dimensions strength, coordination, and flexibility. For the dimension endurance, neither direct nor indirect effects were significant. In the opposite direction, the effects of physical activity on motor abilities were partially mediated by the self-concept of strength. For the dimensions endurance, coordination and flexibility, only indirect were significant. The results of this study support the assumption that the relationship between motor abilities and physical activity is mediated by physical self-concept in both directions. Physical self-concept seems to be an important determinant of adolescents´ physical activity.