Background <p>While parenting style is a well-established correlate of adolescent aggression, prior research has rarely compared specific parenting behaviours across distinct proactive and reactive aggression profiles. The present study addressed this gap by integrating specific parenting behaviours and proactive and reactive aggression within both subgroup-based and variable-centred analytic frameworks.</p> Methods <p>A total of 1,300 Chinese adolescents and their parents from four middle schools in Hong Kong completed questionnaires. Adolescents were classified into non-aggressive, reactive, proactive, and co-occurring reactive-proactive groups using a composite classification based on standardized parent- and self-reported aggression scores. Parenting was assessed at both the global style level (authoritative, authoritarian, and permissive) and the behaviour level (warmth and acceptance, reasoning and induction, democratic participation, physical coercion, verbal hostility, and non-reasoning or punitive discipline). Subgroup-based multivariate analyses of covariance were conducted to compare parenting patterns across aggression groups, followed by hierarchical regression models predicting proactive and reactive aggression from parenting variables.</p> Results <p>Subgroup-based analyses showed that, compared with non-aggressive adolescents, aggression groups generally experienced more maladaptive parenting patterns, especially higher authoritarian and permissive parenting. The proactive group reported the lowest levels of positive parenting, whereas the co-occurring reactive-proactive group showed the highest mean levels on several negative parenting dimensions, although overlap among aggression groups remained substantial. Variable-centred analyses further showed that parenting explained substantially more variance in parent-reported than in self-reported aggression. Across parent-reported models, permissive parenting, physical coercion, and non-reasoning or punitive discipline emerged as the most consistent positive correlates of both reactive and proactive aggression.</p> Conclusions <p>Together, these findings suggest that parenting-aggression associations are overlapping but not uniform across adolescent aggression profiles. Compared with broad parenting styles, behaviour-level indicators provided a more specific account of the parenting practices associated with aggression. These results identify potentially modifiable family correlates that may help guide future longitudinal and intervention research on adolescent aggression.</p>