Purpose This study aimed to assess the role of refugee status and specific mental disorders regarding subsequent labour market marginalisation. Methods Prospective cohort study of all refugees (n = 216,930) and Swedish-borns (n = 3,841,788), aged 19-60 years, and resident in Sweden in 2009. Hazard ratios (HRs) with 95% Confidence Intervals (CIs) for long-term unemployment (> 180 days) and disability pension (DP) were calculated with Cox regression analyses. Results Mental disorders were more prevalent in refugees compared to Swedish-born individuals, with greatest differences seen for post-traumatic stress disorder (PTSD; refugees 1.3%; Swedish-born individuals 0.1%). Regarding long-term unemployment, refugees without a mental disorder had an adjusted HR (aHR) of 2.68 (95% CI 2.65-2.71) compared to Swedish-born individuals without mental disorders, which was above the aHR of refugees (aHR 2.33, 95% CI 2.29-2.38) and Swedish-born individuals (aHR 1.44, 95% CI 1.43-1.45) with mental disorders. Regarding DP, compared to Swedish-born individuals without mental disorders, the aHRs were 1.44 (95% CI 1.34-1.54) for refugees without, but 6.11 (95% CI 5.84-6.39) for refugees with mental disorders. Swedish-born individuals with mental disorder had an aHR of 3.96 (95% CI 3.85-4.07). With regard to specific disorders, the aHRs for refugees, as compared to Swedish-born individuals without mental disorders, were markedly increased for all disorders (e.g. PTSD: long-term unemployment aHR: 2.03 (95% CI 1.89-2.18); DP 7.07 (95% CI 6.42-7.78). Conclusion Mental disorders are more prevalent in refugees than in Swedish-born individuals but do not appear to increase their risk of long-term unemployment. Refugee status and mental disorders strongly contribute to the risk of DP, indicating that factors beyond medical considerations contribute to their granting of DP.