Sickness absence certified with psychiatric diagnoses is increasing in many Western countries. A substantial proportion of the sickness absentees never return to work, but ends up with a permanent disability pension (DP). This study investigated the incidence of long-term sickness absence (LTSA) with different psychiatric diagnoses, and tested predictors of the transition to permanent DP. A special objective was to explore previously reported gender differences, using a population-based cohort study. The population at risk of LTSA was 106,674 occupational active men and 89,356 women in a Norwegian county in 1994; 314 women and 203 men with LTSA >8 weeks, certified with a psychiatric diagnosis, were followed to the end of 1999, with DP as the endpoint. Diagnoses on sickness certificates, age, gender and income were used as explanatory variables in Cox regression analysis. Annual incidence of LTSA with a psychiatric diagnosis was 7.0/1000/year for women and 3.8/1000 for men; 72% of the women and 50% of the men had a diagnosis indicating depression. During follow-up, 32% of the men and 25% of the women obtained DP. Increasing age, male gender, low income and a diagnosis of psychosis or "other" increased the DP risk. Separate analysis for men indicated a different effect of age and a larger role of serious mental illness. The study verified that women more often than men had LTSA with psychiatric diagnoses, especially with depression. Men were at higher risk of transition to DP, and the study suggested some possible explanations.