Our objective was to investigate the associations of vitamin D serum levels with dementia and cognitive function in specific domains in community dwelling older adults. Between 2009 and 2010, we conducted a cross-sectional study in 1,373 individuals (56% men) aged 65+ years in the "Activity and Function in the Elderly in Ulm" (ActiFE) study. Dementia was defined as a Mini-Mental State Examination (MMSE) score ≤ 24. The 25-OHD serum level [ng/mL] was measured by an electrochemilumineszenz immunoassay (ECLIA). Logistic regression models were used to calculate odds ratios (OR)s for cognitive domains (cut-point: 10th percentile) by serum 25-OHD concentrations (both continuously and by cut-point of 20 ng/ml for vitamin D deficiency). Mean age of the study population was 75.6 (SD 6.6) years.We identified 75 participants (43% women) with dementia. 25-OHD concentrations were significantly lower in the participants with dementia compared to persons with a MMSE score >24. We also observed an association of continuous 25-OHD serum concentrations with prevalence of dementia (crude OR 1.05, 95% confidence interval (CI), 1.01-1.08, p-value 0.009) per 1 ng/mL decrease, after adjustment the OR was 1.03, 95% CI, 0.995-1.08 (p-value 0.09). Although vitamin D deficiency was tentatively associated with severity of dementia measured by MMSE (OR 1.35, 95% CI, 0.84-2.19), the association was not statistically significant. However, deficits in specific cognitive domains such as executive functions, wordlist encoding, and visual memory (encoding and recall) were significantly associated with low vitamin D concentration. Our results suggest an association between vitamin D deficiency and cognitive function in specific domains in community dwelling older adults.