The relationship between schizophrenia and violence has been well-established. Yet very little prior research exists on the factors that might explain the nature of this relationship and even fewer studies seek to clarify the etiology of aggressive behavior in adolescents with specific features of schizotypal personality that might help improve the specificity of intervention. The current study tested whether one dimension of schizotypy alone (i.e., the 'suspicious' feature) or the other 8 dimensions (i.e., the 'non-suspicious' features) were particularly associated with aggressive behaviors (reactive and proactive aggression), and if peer problems and low self-esteem mediated these relationships. A serial multiple mediation model testing the hypothesized flow from suspicious and non-suspicious schizotypy to peer problems to low self-esteem and to increased aggression was tested in Hong Kong schoolchildren aged 8- to 14-years (N = 1412; M = 11.47, SD = 1.67 years, female = 47.6%). Increased suspicious and non-suspicious schizotypal features were found to be independently associated with increased reactive aggression, but not proactive aggression. Children with high levels of suspicious schizotypy and non-suspicious schizotypy were more likely to have poor peer problems and low self-esteem concurrently, which in turn was associated with reactive aggression only. This explanatory model suggests that future longitudinal intervention studies that enhance self-esteem in schizotypal adolescents may potentially reduce co-morbid reactive aggressive behaviors too.