<p>With the increasing prominence of adolescent mental health issues, depression has become a significant factor affecting the physical and psychological development of high school students. This study aims to explore the latent classes of depressive symptoms among Chinese high school students and to analyze the influence of various psychological variables on different classes of depression using the emotion regulation framework.&#xa0;The study sampled 1,308 high school students and employed latent class analysis (LCA) to classify their depressive symptoms. Subsequently, six machine learning algorithms were used to analyze the main factors influencing the different depressive categories.&#xa0;The LCA results indicated that the optimal model consisted of three depressive symptom groups: low, moderate, and high depressive symptoms. The random forest algorithm performed the best, with accuracy of 81.02%. Feature importance analysis revealed that core self-evaluation, trait mindfulness, negative emotional regulation efficacy, friendship quality, and parent-child relationship had significant impacts across different depressive categories.&#xa0;Therefore, depressive symptoms in high school students can be categorized into distinct latent classes, and machine learning algorithms effectively identify both the categories and their influencing factors. The findings provide scientific support for the early identification of depression and the development of personalized intervention strategies.</p>