Background: The fact that Cannabis use can lead to psychotic symptoms has been recognized years ago. Moreover, when the patient is using cannabis, the distinction between a primary psychotic disorder and a cannabis-induced psychosis seems to be critical for the prognosis. However, few studies focused on the differences between these two diagnostic groups. We hypothesize that cannabis-induced psychotic disorders have particular demographic, premorbid and clinical features. Our objective was to identify main factors associated to cannabis-induced disorders. Methods: We conducted a retrospective study for a period of twelve years, from January 2002 to December 2013. The study sample was composed of patients hospitalized in the psychiatric department, who reported cannabis-use, and those whose blood and urinary toxicological screening have shown cannabis use. Demographic, family and clinical features were assessed. Psychotic disorders were diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition criteria (DSM-IV), and the study population was divided into two groups, according to causality link of the cannabis on the psychotic disorders. Findings: The sample was composed of 75 patients. Overall, 53 (70.66%) were diagnosed as having a primary psychosis, and 22 (29.33%) were diagnosed as having cannabisinduced psychosis. Significant differences were observed in two domains, concerning family and clinical features. The multivariate data analysis using logistic regression has shown four predictors as being greater in the cannabisinduced psychosis group. The first factor was the age below 25 years old. Subjects in the induced psychosis group were younger, having a median age of 25.1 years compared with 32.1 years for subjects in the cannabis-induced psychosis. The second factor was the marital status. Single or separated subjects were developing more cannabis-induced disorders (Odds Ratio (OR), 2.5; 95% Confidence Interval (CI), 0.69-8.96 the Pearson’s correlation factor (p 0.09). The third factor was family history of psychiatric disorders ((OR), 2.6; 95% CI, 1.14-5.9). The last key factor was an early age on the exposure to the substance, below age of 25 years old (0.02). Conclusion: Differences between substance-induced and comorbide substance-use disorders permit identification of predictors of a substance-induced psychosis. Those factors could help clinicians to classify correctly an early-phase psychotic disorder that co-occurs with substance use in order to challenge treatment and manage severe and persistent mental disorders.