Objectives Job demands and control at work and their combination, job strain, have been studied in relation to risk of disability pension (DP) previously. In the present study, based on registry data, we aimed to deepen the knowledge by analyzing major disease groups among the DPs, dose–response shape of the associations, and potential confounding effects of physical workload.

Methods Approximately 1.8 million workers aged 44 or older and living in Sweden in 2005 were followed up for 16 years, up to a maximum of 65 years of age. We linked mean values of job demands and job control, estimated in a job-exposure matrice (JEM) by gender, to individuals through their occupational titles in 2005. These values were categorized by rank order, and, for the construction of job-strain quadrants, we used a median cut-off. Associations with DP were estimated in Cox proportional-hazards models.

Results In models accounting for covariates including physical workload, low levels of job control were associated with higher risk of DP among both men and women. This association was most clear for DP with a psychiatric diagnosis, although a dose–response shape was found only among the men. High levels of job demands were associated with decreased risk of DP across diagnoses among men, but the same association varied from weak to non-existing among women. The high- and passive job-strain quadrants both showed increased risk of DP with a psychiatric diagnosis.

Conclusion The results suggest that, at the occupational level, low job control, but not high job demands, contributes to an increased incidence of DP, particularly regarding DP with a psychiatric diagnosis.