This study tested the utility of three different models of personality, namely the social and personality model, the pathological personality traits model, and the psychological dysregulation model, in predicting overt aggression, relational aggression, and delinquency in a sample of detained boys (ages 12 to 18; M age = 15.31; SD = 1.16). Results indicated that the three personality approaches demonstrated different unique associations with aggression and delinquency. The psychological dysregulation approach, composed of behavioral dysregulation, emotional dysregulation, and cognitive dysregulation, emerged as the overall best predictor of overt aggression, relational aggression, and delinquency. After controlling for the Big Five personality traits, psychological dysregulation accounted for significant variance in overt aggression and delinquency, but not relational aggression. After controlling for callous-unemotional traits and narcissistic traits, psychological dysregulation also accounted for significant variance in overt aggression, relational aggression, and delinquency. Psychological dysregulation did not account for significant variance in aggression or delinquency after controlling for borderline traits. The pathological personality traits approach, comprised of callous-unemotional traits, narcissistic traits, and borderline traits performed second best. In particular, within this approach borderline traits accounted for the most unique variance, followed by narcissistic traits, then callous-unemotional traits. Borderline traits accounted for significant variance in overt aggression, relational aggression, and delinquency when controlling for the Big Five traits, but not after controlling for psychological dysregulation. Narcissistic traits only accounted for significant variance in overt aggression and relational aggression after controlling for the Big Five personality traits, but not after controlling for psychological dysregulation. CU traits only accounted for significant variance in overt aggression after controlling for the Big Five personality traits, but not after controlling for psychological dysregulation. The social and personality model, represented by the Big Five personality traits accounted for the least amount of variance in the prediction of aggression and delinquency, on its own, and when pitted against the other two personality approaches. The exception was that the Big Five personality traits accounted for significant variance in relational aggression beyond narcissistic traits, as well as psychological dysregulation. These findings have implications for assessment and intervention with aggressive and antisocial youth.