BACKGROUND: In South Korea, as an aged society, an understanding of dementia and its risk factors is important from clinical and healthcare policy perspectives. Relationship between cognitive impairment and body weight or weight changes have been reported, but these were contradictory. We have evaluated the association between weight changes and cognitive decline using national level longitudinal data. METHODS: Data from the Korean Longitudinal Study of Ageing from 2006 to 2012 were used. Association between weight changes and decline in cognitive function as measured by K-MMSE (the Korean version of the Minimental state examination) score was assessed by multivariate logistic regression. Weight changes were calculated from 1st wave and 3rd wave survey data, and classified into five groups as stable, increases, decreases of >10%, or 5%–10%. RESULTS: About 37% of the total participants (n=4,512) were 65 years or older. These participants made up the largest proportion of the groups with weight change exceeding 10%. Multivariate logistic regression analyses revealed that weight changes exceeding 10% (10% increase vs stable, adjusted OR [aOR] 1.47, 95% confidence interval (CI) 1.11–1.95; 10% decrease vs stable, aOR 1.44, 95% CI 1.11–1.88) were significant predictive factors for decline in cognitive function. In subgroup analyses, the association between weight changes and cognitive decline was significant in males aged over 65 years and in normal BMI groups. CONCLUSION Weight changes, both increases and decreases exceeding 10% of baseline, were significantly associated with declines in cognitive function among older adults in South Korea.