The objectives of this article are to identify groups of trajectories of stimulant use across adolescence and early adulthood among juvenile offenders and to identify risk factors predicting development. Group-based trajectory modeling was used to identify developmental subgroups of stimulant users during the study period. Multinomial logistic regression was used to examine baseline covariates’ relevance for predicting assignment to identified trajectory groups. A six-group model of stimulant use was found to best fit the data. Race, history of direct victimization at baseline, gender, low school commitment at baseline, low religiosity at baseline, and other substance use at baseline were all identified as predictors of trajectory group assignment. Juvenile offenders, particularly White and female offenders, are at elevated risk of illicit stimulant use during adolescence and early adulthood. Despite this, the vast majority of juvenile offenders in the sample had desisted by age 23, differing from research on other substances using this same sample.