BACKGROUND: Incident diabetes risk is inversely proportional to 25-hydroxyvitamin D [25(OH)D] levels among non-Hispanic white but is unclear among African American (AA) populations. Serum 25(OH)D2 may be an important component of total 25(OH)D among AA populations due to higher levels of melanin. OBJECTIVE: To assess the association of serum 25(OH)D with incident diabetes among AAs and stratify by detectable 25(OH)D2. DESIGN: Serum 25(OH)D2 and 25(OH)D3 were collected from 2000 to 2004 among AA participants in the Jackson Heart Study. A cosinor model was used to adjust for the seasonality of 25(OH)D3; 25(OH)D3 and 25(OH)D2 were combined to ascertain total 25(OH)D. Incident diabetes (fasting glucose ≥126 mg/dl, use of diabetes drugs, or HbA1c ≥6.5%) was assessed over 12 years among adults without diabetes at baseline. Participants with missing baseline covariates or diabetes follow-up were excluded. Hazard ratios (HR) were estimated using Cox modeling, adjusting for age, sex, education, occupation, smoking, physical activity, alcohol use, aldosterone, and body-mass index. RESULTS: Among 3311 adults (mean age 53.3 years, 63% female) 584 participants developed diabetes over a median of 7.7 years. After adjustment, 25(OH)D ≥20 compared to <12 ng/ml was associated with a HR 0.78 (95% CI: 0.61, 1.00). Among participants with detectable 25(OH)D2 and 25(OH)D3 (n = 1671), 25(OH)D ≥ 20 ng/ml compared to <12 ng/ml was associated with a 35% (HR 0.65, 95% CI: 0.46, 0.91) lower risk of diabetes. CONCLUSIONS: Higher levels of 25(OH)D may be protective against the development of diabetes among AA individuals, particularly among those with detectable 25(OH)D2 and 25(OH)D3.