Growth mixture modeling was used to identify different trajectories of body mass index (BMI) among adolescents ages 10–15 from a national sample. Three distinct classes were found for both boys and girls: “normative” (90.9% and 89.7%), “high increasing” (6.3% and 7.4%), and “decreasing” (2.8% and 2.9%). Multinomial logistic regression identified family income as predictive of class membership for boys and pubertal status and being rural as predictive for girls. Parent‐reported health was a common predictor across gender. Growth curves of internalizing symptoms and physical activity were modeled to explore trends across classes. Findings highlight complexities in the relations between BMI, internalizing symptoms, and physical activity in this developmental period.