Diabetes has been associated with increased risk of endometrial cancer in some epidemiological studies. Body mass index (BMI) and other measures of obesity have been associated positively with both diabetes and endometrial cancer. It is not clear whether or not the association of diabetes with endometrial cancer is explained entirely by obesity. Thus, we sought to test the hypothesis that diabetes is not associated with endometrial cancer independent of obesity. We examined the association between self-reported diabetes (onset at >30 years of age) and incident endometrial cancer in a prospective cohort study of 24,664 postmenopausal women in Iowa. Over 12 years of follow-up, 346 cases occurred among the cohort at risk. Data were analyzed using proportional hazards regression models. Diabetes was analyzed as reported at baseline and as a time-dependent variable using information obtained during follow-up. After adjustment for BMI, waist:hip ratio, and other covariates, the relative risk (RR) for women with diabetes versus women without diabetes was 1.43 [95% confidence interval (CI), 0.98-2.1]. The diabetes association was confined to women in the upper two BMI quintiles (RR, 1.47; 95% CI, 0.98-2.20), but a formal test of interaction was not statistically significant. Analyses that included diabetes ascertained at baseline and at follow-up gave similar results; the diabetes-associated RR in the higher BMI strata was 1.64 (95% CI, 1.16-2.31). We conclude that after adjustment for other risk factors, diabetes is associated with a modestly increased risk for endometrial cancer among women in this cohort.