Our study's objective is to determine whether substance use disorders’ association with aggression differs according to the type of substance and/or the form of aggression, within the same population. We used data from the National Survey on Drug Use and Health across 2008–2014, with a pooled sample of 270,227 adult respondents. We used regression models to estimate the odds ratios for those having alcohol and/or drug use disorder(s) perpetrating (a) each form of aggression compared with no aggression and (b) other‐directed compared with self‐directed aggression. Alcohol use disorder alone and drug use disorder(s) alone were both associated with significantly increased odds of committing self‐directed, other‐directed, and combined aggression. Individuals with drug use disorder(s) alone were more likely to commit other‐directed than self‐directed aggression (adjusted odds ratio = 1.46, 95% CI = 1.04–2.05). Individuals with alcohol use disorder alone were not likely to commit one over the other (adjusted odds ratio = 1.20, 95% CI = 0.90–1.61). In conclusion, the integrated model of aggression based on the stress–diathesis model is a relevant framework to study risk factors for aggression. Further research is needed to identify longitudinal predictors of directionality of aggression.