Purpose: The aim of this study was to test the hypothesized structural paths in Bandura's social-cognitive theory (SCT) model on adolescent girls' physical activity following a 12-month physical activity and dietary intervention to prevent obesity. Method: We conducted a 12-month follow-up study of 235 adolescent girls (M^sub age^ = 13.2 years, SD = 0.4) from 12 secondary schools located in low-income communities. At baseline, participants completed SCT scales related to physical activity (i.e., self-efficacy, intention, parental support, and outcome expectations). At baseline and 12-month follow-up (postintervention), participants wore accelerometers for 7 days. Structural equation modeling was used to determine if Time 1 measures predicted physical activity at 12-month follow-up after adjusting for baseline activity. Results: The model explained 28% and 34% of the variance in physical activity and intention, respectively. Model fit indexes indicated the data were a good fit to the model; however, only self-efficacy was associated with physical activity at 12 months. There was no support for intention or outcome expectations as proximal determinants of behavior. Self-efficacy was associated with outcome expectations and parental support; however, only outcome expectations predicted intention. Conclusions: Current findings indicate a large proportion of the variance for physical activity and intention remains unexplained and that the proposed pathways in the SCT model were not fully supported. Future model testing may need to consider augmentation or integration of theoretical models, which may include ecological components if we are to advance our understanding of physical activity behavior in this subgroup of the adolescent population.