OBJECTIVE - To investigate the relationship between diabetes and future risk of Parkinson's disease (PD) among older U.S. adults. RESEARCH DESIGN AND METHODS - A prospective study of self-reported diabetes in 1995 and 1996 in relation to PD diagnosed after 1995 among 288,662 participants of the National Institutes of Health-AARP Diet and Health Study. Multivariate odds ratio (OR) and 95% CI were derived from logistic regression models. RESULTS - A total of 1,565 participants with PD diagnosed after 1995 were included in the analysis. After adjustment for potential confounders, PD risk was ∼40% higher (OR = 1.41 [95% CI 1.20-1.66]) among diabetic patients than among participants without diabetes. Further analysis showed that the risk elevation was largely limited to individuals who had diabetes for more than 10 years at the time of baseline survey (1.75 [1.36-2.25]). The association with diabetes was seen for both participants with PD diagnosed between 1995 and 1999 and participants with PD diagnosed after 2000. In addition, similar results were obtained after excluding participants with stroke, heart disease, cancers, or poor or fair health status and in subgroup analyses by age, sex, smoking status, and coffee consumption. CONCLUSIONS - This large study showed that diabetes was associated with a higher future risk of PD and the nature of this association warrants further investigation.