Abstract Background: Although evidence from psychosis patients demonstrates the adverse effects of cannabis use (CU) at a young age and that the rate of CU is high in subgroups of young violent patients with psychotic disorders, little is known about the possible effect of the age of onset of CU on later violent behaviors (VB). So, we aimed to explore the impact of age at onset of CU on the risk of displaying VB in a cohort of early psychosis patients. Method: Data were collected prospectively over a 36-month period in the context of an early psychosis cohort study. A total of 265 patients, aged 18–35 years, were included in the study. Logistic regression was performed to assess the link between age of onset of substance use and VB. Results: Among the 265 patients, 72 had displayed VB and 193 had not. While violent patients began using cannabis on average at age 15.29 (0.45), nonviolent patients had started on average at age 16.97 (0.35) (p = 0.004). Early-onset CU (up to age 15) was a risk factor for VB (odds ratio = 4.47, confidence interval [CI]: 1.13–20.06) when the model was adjusted for age group, other types of substance use, being a user or a nonuser and various violence risk factors and covariates. History of violence and early CU (until 15) were the two main risk factors for VB. Conclusions: Our results suggest that early-onset CU may play a role in the emergence of VB in early psychosis.