Background: Moderate-to-vigorous physical activity (MVPA) declines during adolescence, particularly among girls. In this randomized control trial, we examined MVPA in a physical activity intervention compared to physical education (PE) class as usual (TAU), stratified by sex and weight classification. Standardized BMI (zBMI) overtime was also examined. Methods: Hispanic-American middle school students (N = 193) were recruited from a school district in Houston, Texas. Participants were randomized to either a circuit-based physical activity intervention or TAU (PE class as usual). MVPA was assessed using accelerometry at baseline and 6 months. Repeated measures ANCOVA were conducted to examine changes in MVPA, overall and stratified by sex and weight classification. This procedure was repeated for zBMI. Results: Participants were 12.10 [plus or minus] 0.63 years old and 53% were girls. Overall those in intervention increased weekday MVPA more than TAU (F(1,190) = 7.03, p < 0.01). Intervention girls increased weekday MVPA; whereas TAU girls decreased weekday MVPA (F(1,99) = 7.36, p < 0.01). Among those with obesity, there was no difference in MVPA between conditions (F(1, 56) = 0.33, p = 0.57), but Intervention decreased zBMI significantly more than TAU (F(1, 56) = 6.16, p < 0.05). Conclusions: Structured PE classes grounded in behavioral theory may be an important strategy to prevent typical decreases in MVPA during adolescence, particularly among girls and for youth with obesity.