PURPOSE: Little is known about the relationship between social exclusion and cognitive impairment in psychosis. We conducted a long-term cohort study of first-episode psychosis to examine the association between comprehensive measures of cognitive impairment and social exclusion assessed at follow-up. METHODS: A total of 173 subjects with first-episode psychosis were assessed after a 20-year follow-up for 7 cognitive domains and 12 social exclusion indicators. Associations between sets of variables were modeled using multivariate regression, where social exclusion indicators were the dependent variables, cognitive domains were the independent variables, and age, gender, and duration of follow-up were covariates. RESULTS: The total scores on the measures of cognition and social exclusion were strongly associated (β = - .469, ∆R CONCLUSION: The domains of cognitive impairment were strongly and differentially related to those of social exclusion. Given that such an association pattern is likely bidirectional, a combined approach, both social and cognitive, is of paramount relevance in addressing the social exclusion experienced by individuals with psychotic disorders.