Background Neuroimaging methods that allow researchers to investigate structural covariance between brain regions are increasingly being used to study psychiatric disorders. Structural covariance analyses are particularly well suited for studying disorders with putative neurodevelopmental origins as they appear sensitive to changes in the synchronized maturation of different brain regions. We assessed interregional correlations in cortical thickness as a measure of structural covariance, and applied this method to investigate the coordinated development of different brain regions in conduct disorder (CD). We also assessed whether structural covariance measures could differentiate between the childhood‐onset (CO‐CD) and adolescence‐onset (AO‐CD) subtypes of CD, which may differ in terms of etiology and adult outcomes. Methods We examined interregional correlations in cortical thickness in male youths with CO‐CD or AO‐CD relative to healthy controls (HCs) in two independent datasets. The age range in the Cambridge sample was 16–21 years (mean: 18.0), whereas the age range of the Southampton sample was 13–18 years (mean: 16.7). We used FreeSurfer to perform segmentations and applied structural covariance methods to the resulting parcellations. Results In both samples, CO‐CD participants displayed a strikingly higher number of significant cross‐cortical correlations compared to HC or AO‐CD participants, whereas AO‐CD participants presented fewer significant correlations than HCs. Group differences in the strength of the interregional correlations were observed in both samples, and each set of results remained significant when controlling for IQ and comorbid attention‐deficit/hyperactivity disorder symptoms. Conclusions This study provides new evidence for quantitative differences in structural brain organization between the CO‐CD and AO‐CD subtypes, and supports the hypothesis that both subtypes of CD have neurodevelopmental origins.