While an association between physical activity and ovarian cancer is biologically plausible, results from previous studies have been inconsistent. This relationship was investigated using data collected from 240 cases and 891 controls in a population-based case-control study.Strenuous recreational activity in the mid-teens and the early thirties was associated with an increased risk of ovarian cancer, while strenuous and moderate recreational activity and increasing intensity of occupational activity two years prior to questionnaire completion was associated with a decreased risk. A non-significant reduction in risk was found for occupational activity in the early fifties. BMI was not a significant effect modifier in all analyses. Exclusion of borderline tumours did not significantly alter results, although increased risk was seen for moderate activity during the mid-teens. Strenuous recreational activity early in life may increase the risk of ovarian cancer, while higher intensity occupational activity and more recent recreational activity reduces the risk.