Abstract Background: Several observational studies assessed the relationship between serum 25-hydroxyvitamin D [25(OH)D] concentrations and the risk of cancer but results were inconclusive. Methods: We measured 25(OH)D concentrations in a population-based cohort study of 9,949 men and women ages 50 to 74 years in Saarland, Germany. Comprehensively adjusted Cox regression models were applied to estimate HRs and 95% confidence intervals (CI) for the association between season-standardized 25(OH)D concentrations and total and site-specific cancer incidence. Results: Overall, during a median of 8 years of follow-up, 873 subjects developed cancer; the most common being prostate (171), breast (137), lung (136), and colorectal (136) cancer. Low season-standardized 25(OH)D (<30, 35, 40, or 36 nmol/L in winter, spring, summer, and autumn, respectively) was neither significantly associated with total cancer incidence (HR, 1.10; 95% CI, 0.93–1.30) nor with site-specific cancer incidence. However, a significantly increased overall cancer risk was observed for low 25(OH)D among men, nonobese subjects and subjects reporting low fish consumption and for high 25(OH)D in nonsmokers and nonobese subjects. Accordingly, restricted cubic splines to investigate dose–response relationships curves showed an inverse association of 25(OH)D levels and total cancer risk in men but not in women. Conclusions: 25(OH)D concentrations were significantly associated with overall cancer incidence in subgroups of this large cohort from Germany. No significant association was observed with site-specific cancers but this could be due to a limited statistical power for these endpoints. Impact: Further research should clarify whether and to what extent specific risk groups might profit from vitamin D supplementation. Cancer Epidemiol Biomarkers Prev; 22(5); 905–16. ©2013 AACR.