Aims: We aimed to examine how newly diagnosed diabetes and work disability jointly predict death during working age. Methods: We used prospective population-based register data of 25–59-year-old adults who had lived in Sweden since 2002. All those with onset of diabetes recorded in 2006 were included (n=14266). A 2% random sample (n=78598) was drawn from the general population, comprising people with no indication of diabetes during 2003–2010. Net days of sickness absence and disability pension in 2005–2006 were examined; the follow-up time for mortality was 2007–2010. Cox regression models were fitted (hazard ratios, HR, 95% confidence interval, CI) adjusting for sociodemographics and time-dependent health conditions. Results: Individuals with diabetes and work disability for over 6 months were at a higher risk of premature death (HR=14.2, 95% CI 12.0–16.8) than their counterparts without diabetes and work disability. A high risk was also observed among people without diabetes but equally prolonged work disability (HR=6.4, 95% CI 5.4–7.6). Diabetes was associated with premature death even without work disability (HR=3.5, 95% CI 2.8–4.4). The associations were particularly attenuated after adjustment for health conditions assessed over the follow-up. Conclusions: Diabetes and work disability jointly increase the risk of death during working age. Diabetes with long-term work disability is associated with the highest risk of premature death, which highlights the importance of their prevention and early detection.