<p>To explore the impact of emotion regulation on adolescents' emotional health and the protective mechanism of physical exercise, this study selected 5423 students (2749 middle school students and 2674 university students) as research subjects. Latent profile analysis was employed to explore the combination model of emotion regulation strategies and emotion regulation flexibility, as well as the mediating role between physical exercise and depression and anxiety. The findings revealed the following: (1) Early adolescents (middle school group) and late adolescents (university group) exhibited different emotion regulation patterns, which manifested in different numbers of latent subgroups and proportions of individuals. There are two latent subgroups of emotion regulation in the early adolescent group: the "LowCR-LowERF" and "HigCR-HigERF". The number of individuals is mainly concentrated in the LowCR-LowERF (accounting for 69%). There are four latent subgroups of emotion regulation in the late adolescent group: the "LowCR-LowERF", the "HigRS-LowERF", the "HigCR-HigERF", and the "MidCR-MidERF". The number of individuals is mainly concentrated in the "LowCR-LowERF" (accounting for 41%). (2) Compared to emotion regulation strategies, emotion regulation flexibility had a stronger predictive effect on adolescents' emotional health. Individuals with higher emotion regulation flexibility tended to experience less depression and anxiety and more positive emotions. (3) Physical exercise significantly predicted the emotion regulation subgroups in both early and late adolescents, and it served as a protective factor for emotional health. (4) The mediating role of emotion regulation strategies and flexibility was significant. The importance of physical exercise in enhancing adolescents' emotion regulation flexibility and improving negative emotions should be emphasized.</p>