The association between disturbed sleep and increased risk of occupational injury has been observed in several cross-sectional and case-control studies, but prospective evidence is lacking. We examined prospectively whether sleep disturbances predicted occupational injuries in a large population of Finnish public sector employees. A total of 48 598 employees working in 10 municipalities and 21 hospitals in various parts of Finland were included. Sleep disturbances were assessed with the four-item Jenkins Sleep Problems Scale. Records of sickness absence due to occupational injury during the year following the survey were obtained from employers' registers. A proportion (9076; 22%) of participants reported disturbed sleep, and 978 (2.4%) had a recorded occupational injury. After adjustment for socio-demographic characteristics, the odds ratio (OR) for occupational injury was 1.38 [95% confidence interval (CI) 1.02-1.87] times higher for men with experiences of disturbed sleep than for those without sleep disturbances, but not significant for women. Of the sub-dimensions of sleep disturbances, the OR for occupational injury was 1.69 (95% CI 1.26-2.26) for women with difficulties initiating sleep, but not significant for men. These associations remained after additional adjustment for work stress, sleep length, obesity, alcohol use and mental health. This study suggests that sleep disturbances are a significant predictor of occupational injuries even after accounting for a range of covariates.