•The polygenic risk score for educational attainment (PRS-EA), rather than psychopathological PRS (schizophrenia or bipolar disorder), was associated with lower risk and a later onset of relapse.•These results could indicate that the biological mechanisms beyond the risk of suffering a psychotic episode and the mechanisms of its prognosis could be independent.•The combination of polygenic scores and clinical data are potentially useful predictors of relapse. Little is known about genetic predisposition to relapse. Previous studies have linked cognitive and psychopathological (mainly schizophrenia and bipolar disorder) polygenic risk scores (PRS) with clinical manifestations of the disease. This study aims to explore the potential role of PRS from major mental disorders and cognition on schizophrenia relapse. 114 patients recruited in the 2EPs Project were included (56 patients who had not experienced relapse after 3 years of enrollment and 58 patients who relapsed during the 3-year follow-up). PRS for schizophrenia (PRS-SZ), bipolar disorder (PRS-BD), education attainment (PRS-EA) and cognitive performance (PRS-CP) were used to assess the genetic risk of schizophrenia relapse.Patients with higher PRS-EA, showed both a lower risk (OR=0.29, 95% CI 0.11–0.73) and a later onset of relapse (30.96± 1.74 vs. 23.12± 1.14 months, p=0.007. Our study provides evidence that the genetic burden of neurocognitive function is a potentially predictors of relapse that could be incorporated into future risk prediction models. Moreover, appropriate treatments for cognitive symptoms appear to be important for improving the long-term clinical outcome of relapse.