Background: The increasing aging of the population with the consequent increase of age-associated cognitive disorders pose the challenge of controlling its preventable risk factors, among which vitamin D deficit is a putative factor. Thus, our objective is to explore the association between vitamin D and cognitive performance in a cohort study of community-dwelling Chilean older people. Material and Methods: Cohort study of 955 (69.7% female), community-dwelling older Chileans free of cognitive impairment from the Alexandros cohorts, with 25(OH)D measurement at baseline. Cognitive Function was evaluated with the Mini Mental State Examination (MMSE) short-form questionnaire. Plasma levels of 25(OH)D were classified as Normal > 30 ng/mL Insufficiency 20-29 ng/mL, Deficiency 20-12 ng/mL and Severe Deficiency < 12 ng/mL. Penalized regressions models were made to assess associations. Results: Mean age of the sample was 66.6 + 4.5 years, with 8.5 + 4.7 years of education. After a mean follow-up of 9.6 years, 54 new cases of Mild Cognitive Impairment (MCI)were identified (Incidence density rate = 5.9 per 1000 person/years). Mean vitamin D plasma levels were lower in people with MCI than in the normal cognitive ones (23.0 + 12.75 vs. 28.35 + 15.17 ng/mL, p < 0.01). In the fully adjusted model only severe deficiency of vitamin D was associated with MCI (RR = 2.33; 95% CI: (1.03-5.26). Conclusions: In this longitudinal study, our results confirm that low Vitamin D is a risk factor for MCI, and that people with severe deficiency have more than double the risk of MCI people with normal Vitamin D levels. Considering the high frequency of vitamin D deficiency in older people, and its preventability, these results are very valuable for future public health programmes.