Depression has become increasingly prevalent among adolescents, posing significant challenges for mental health professionals. While most studies on depression adopt a cross-sectional perspective or a variable-centred approach, these methods often fail to illuminate the developmental trajectories of depression in individuals. We employed Latent Profile Transition Analysis (LPTA), a person-centred approach, to analyse longitudinal data from a large adolescent sample (N = 978; Mage = 16.26, SD = 0.89; 52.2% females). This study aimed to identify distinct subgroups of depression and observe transitions between these groups over time, considering stress, anxiety, and rumination as covariates to predict these transitions and aid in the development of targeted interventions. We identified three distinct subgroups: 'low/no depression', 'moderate depression', and 'high depression'. Individuals in the low/no depression and moderate depression groups displayed a predominant tendency toward stability rather than change. Conversely, individuals in the high depression group showed a high probability of transitioning to the moderate depression group. Stress, rumination, and anxiety were significant predictors of transitions into more severe depressive groups. Notably, the predictive power of rumination diminished over time. This study relied solely on self-reported measures, which may introduce response bias. This study reveals dynamic trajectories of depression among adolescents using a person-centred approach, emphasising the importance of closely monitoring those in the moderate depression subgroup. Stress, anxiety, and rumination emerged as crucial predictors of transitions in depression severity, underscoring the need for targeted early interventions.