A medium-to-high level of physical activity (PA) may have at least a short-term positive effect on psychopathology in children and adolescents. Hence, the objective of this study was to investigate the long-term effects of PA in non-adult age groups on their general mental health problems and/or ADHD symptoms, using trajectories of concurrent development over a period of 10 years. This study employed data from the German Health Interview and Examination Survey for Children and Adolescents (KiGGS) collected at three time points (baseline, Wave 1, Wave 2, over 10 years) from 17,640 children and adolescents. Using parent-reported data from the Strengths and Difficulties Questionnaire (SDQ), different developmental trajectories of general mental health problems (SDQ-total) and ADHD symptoms (SDQ-H/I) were identified with latent class mixed models (LCMM) statistics. This was also applied to parent- and self-reported data of three levels of PA. The latter was assessed according to WHO recommendations. The joint probability of class membership for SDQ-total as well as ADHD symptoms with PA was calculated to generate the concurrent developmental trajectories between variables. Results showed a 4-class trajectory model for both SDQ-total and ADHD symptoms among boys and girls. The majority of children and adolescents showed “low general difficulties” and “low ADHD symptoms” over the period of 10 years. Three distinct trajectories in boys and four distinct trajectories in girls were found for PA. Most of the participants showed an “increasing–decreasing activity” trajectory. No statistically significant correlations were found between the different SDQ-total or ADHD symptom trajectories and the trajectories of PA in the two genders. Taken together, our findings did not indicate any significant relationship between waxing and waning PA course over 10 years and various classes of mental health problems for children and adolescents. In contrast to our cross-sectional findings, no steady long-term medium/high-level of PA was present, which could (at least partly) explain the non-significant findings.