Background Most studies on offending heterogeneity have been conducted with general population samples. It is not clear to what extent these can inform such outcomes for individuals with substance use disorders specifically. Aims The aim of this study is to compare the offending trajectories of individuals treated for substance use disorders in adolescence with a matched general population sample, and to test for gender differences in this respect. Method Growth mixture models were applied to identify offending trajectories from age 15 to 33 of 1568 individuals treated for substance use disorders in adolescence, and in a matched population‐based sample of 1500 individuals. Results Several parallel trajectories for men and for women were identified in both samples. The substance misuse treatment sample, however, had higher levels of offending, larger offender classes, longer careers and two additional, distinct trajectories. Although there were similarities between the men and women, the men were more heterogeneous offenders. There were two distinct offending trajectories among male substance misusers—decreasing high level and decreasing low level offending. Conclusions Differences between substance using and general population samples indicate that results from the latter could underestimate the severity, heterogeneity, and persistence of offending trajectories if merely generalised to individuals with substance use disorders. Our results also indicated that population‐based samples might be underpowered for detecting female offending heterogeneity. Copyright © 2014 John Wiley & Sons, Ltd.