We investigated effects of occupational physical activity on relative risk for prostate cancer. From Swedish nationwide censuses in 1960 and 1970, we defined two cohorts of men whose occupational titles allowed classification of physical activity levels at work in 1960 (n=1,348,971) and in 1970 (n=1,377,629). A third cohort included only men whose jobs required a similar level of physical activity in both 1960 and 1970 (n=673,443). The incidence of prostate cancer between 1971 and 1989 was ascertained through record linkage to the Swedish Cancer Register. A total of 43,836, 28,702, and 19,670 prostate cancers, respectively, occurred in the three cohorts. In all three cohorts, the relative risk for prostate cancer increased with decreasing level of occupational physical activity (P<0.001), using Poisson regression. Among men with the same physical activity levels in 1960 and 1970, the rate ratio was 1.11 for men with sedentary jobs as compared with those whose jobs had very high/high activity levels after adjustment for age at follow-up, calendar year of follow-up and place of residence (95% CI 1.05-1.17; P for trend <0.001). There was no association between occupational activity and prostate cancer mortality. Since we had no data on other potential risk factors the observed associations for both incidence and mortality might have been confounded. Further studies are needed to better understand the potential role of physical activity for prostate cancer.