Objectives: We sought to estimate the prevalence of mental illness (MI) and substance use disorders (SUD) and determine the predictors of MI/SUD categories among veteran women with diabetes. Methods: We evaluated a cross-sectional analysis of 16,368 women veterans with diabetes in fiscal 1999 and 2000. SUD, MI, and diabetes were identified using validated algorithms based on diagnosis codes. Chi-square and multinomial logistic regressions were used to examine associations between SUD, MI, demographic, socioeconomic, and health status variables. Results: Overall, 45% of women had a MI, SUD, or both. SUD rates were high among those with serious MI. The associations between MI/SUD and independent variables were not uniformly significant, except for macrovascular conditions; women veterans were more likely to have combinations of MI/SUD. Conclusion: A high rate of MI/SUD suggests that care for mental and physical illness needs to be integrated into health care planning and delivery of services to veteran women with diabetes.