The rapid increase in the prevalence of dementia associated with ageing populations has stimulated interest in identifying modifiable lifestyle factors that could prevent cognitive impairment. One such potential preventive lifestyle factor is the Nordic diet that has been shown to reduce the risk of CVD; however, its effect on cognition has not been studied. The aim of the present study was to estimate the cross-sectional and longitudinal associations of the baseline Nordic diet with cognitive function at baseline and after a 4-year follow-up in a population-based random sample (n 1140 women and men, age 57–78 years) as secondary analyses of the Finnish Dose-Responses to Exercise Training study. The Nordic diet score was created based on reported dietary components in 4-d food records. Cognition was assessed by the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) neuropsychological battery and the Mini-mental State Examination (MMSE). The baseline Nordic diet score had been positively associated with Verbal Fluency (β 0·08 (95 % CI 0·00, 0·16), P= 0·039) and Word List Learning (β 0·06 (95 % CI 0·01, 0·10), P= 0·022) at 4 years but not with the Consortium to Establish a Registry for Alzheimer's Disease total score (CERAD-TS) or MMSE at 4 years, after adjustment for baseline cognitive scores, demographic factors and health-related factors. After excluding individuals with impaired cognition at baseline, the baseline Nordic diet score had also been positively associated with the CERAD-TS (β 0·10 (95 % CI 0·00, 0·20), P= 0·042) and MMSE (β 0·03 (95 % CI 0·00, 0·06), P= 0·039) at 4 years. These associations disappeared after further adjustment for energy intake. In conclusion, the Nordic diet might have a positive association with cognition in individuals with normal cognition.