Background-Several prospective studies have evaluated the association between body mass index (BMI) and death risk among patients with diabetes mellitus; however, the results have been inconsistent. Methods and Results-We performed a prospective cohort study of 19 478 black and 15 354 white patients with type 2 diabetes mellitus. Cox proportional hazards regression models were used to estimate the association of different levels of BMI stratification with all-cause mortality. During a mean follow-up of 8.7 years, 4042 deaths were identified. The multivariable-adjusted (age, sex, smoking, income, and type of insurance) hazard ratios for all-cause mortality associated with BMI levels (18.5-22.9, 23-24.9, 25-29.9, 30-34.9 [reference group], 35-39.9, and >= 40 kg/m(2)) at baseline were 2.12 (95% confidence interval [CI], 1.80-2.49), 1.74 (95% CI, 1.46-2.07), 1.23 (95% CI, 1.08-1.41), 1.00, 1.19 (95% CI, 1.03-1.39), and 1.23 (95% CI, 1.05-1.43) for blacks and 1.70 (95% CI, 1.42-2.04), 1.51 (95% CI, 1.27-1.80), 1.07 (95% CI, 0.94-1.21), 1.00, 1.07 (95% CI, 0.93-1.23), and 1.20 (95% CI, 1.05-1.38) for whites, respectively. When stratified by age, smoking status, patient type, or the use of antidiabetic drugs, a U-shaped association was still present. When BMI was included in the Cox model as a time-dependent variable, the U-shaped association of BMI with all-cause mortality risk did not change. Conclusions-The present study indicated a U-shaped association of BMI with all-cause mortality risk among black and white patients with type 2 diabetes mellitus. A significantly increased risk of all-cause mortality was observed among blacks with BMI <30 kg/m(2) and >= 35 kg/m(2) and among whites with BMI <25 kg/m(2) and = 40 kg/m(2) compared with patients with BMI of 30 to 34.9 kg/m(2).