Background Adults with bipolar disorder (BD) have higher rates of substance use disorders (SUDs) compared to the general population. SUD rates in young offspring/relatives of BD probands, as well as factors which drive those rates, are not as well-characterized. Methods We aimed to examine SUD prevalence among adolescent/young adult offspring and relatives of probands with and without BD. Data were collected from five sites in the US and Australia during 2006–2011. Youth offspring/relatives (“Relatives of BD probands;” n = 267; mean age = 16.8 years; ± 2.9 S.D.), identified through a proband family member with DSM-IV BD (Type I or II), were compared to offspring/relatives of control probands (“relatives of control probands;” n = 149; mean age = 17.4 years; ± 2.9 S.D.). Logistic regression with generalized estimating equations was used to compare the groups across a range of substance use and SUD variables. Odds ratios were calculated for lifetime prevalence of substance outcomes. Results Bivariate analyses showed DSM-IV SUDs were more prevalent among relatives of BD probands than among relatives of control probands (29% vs. 18%; p = 0.01). Generalized estimating equation models showed BD mood and childhood-onset externalizing disorders in adolescent and young adult relatives to each significantly increase the odds (OR = 2.80–3.17; p < 0.02) for the development of several substance variables among all relatives, whereas the risk of SUDs in relatives was not increased when the relatives had no mood or externalizing disorders themselves. Conclusion Relatives of BD probands with lifetime mood and externalizing disorders report more substance use/SUDs than relatives of control probands. In contrast, SUD outcomes in relatives of BD probands without mood or externalizing disorders were no different from control relatives without psychopathology. Early recognition and treatment of psychiatric disorders may lead to less substance use in this highly vulnerable population.