This study describes qualitatively distinct trajectories of BMI change among girls participating in a longitudinal study of non-Hispanic, white girls (n = 182) and their parents, assessed at daughters' ages 5, 7, 9, 11, 13, and 15 years. Height, weight, body fat, fasting blood glucose and lipids, blood pressure, waist circumference, and pubertal status were measured, and participants self-reported dietary, physical activity, and television (TV) viewing patterns. Growth mixture models were used to model heterogeneity in girls' BMI trajectories over 10 years. Statistical support was strongest for four distinct BMI trajectories: (i) upward percentile crossing (UPC; n = 25, 14%); (ii) delayed downward percentile crossing (DDPC; n = 37, 20%); (iii) 60th percentile tracking (60PT; n = 52, 29%); and (iv) 50th percentile tracking (50PT; n = 68, 37%). Girls in the UPC group had more metabolic risk factors at age 15 years, even after adjusting for concurrent weight status. Girls in the UPC group had mothers with the highest BMIs at study entry and were breast-fed for a shorter duration. This novel approach for examining differences in growth trajectories revealed four distinct BMI trajectories that predicted adolescent metabolic health outcomes in girls. The present study provides support for BMI monitoring in girls and for the potential utility of combining data on BMI tracking with data on familial characteristics for the early identification of girls at elevated risk for obesity and metabolic syndrome.