Objectives: This study compares educational differences in the functional limitations of 55-65-year-olds in the Netherlands in 1992 and 2002 and examines whether changes are explained by cohort lifestyle and psychosocial changes. Methods: Data from two cohorts of 55-65-year-olds (n = 948 in 1992 and n = 980 in 2002) in the Longitudinal Aging Study Amsterdam are analysed. Results: Men's disability ratios are similar in both cohorts. The women's disability ratio is higher in 2002 than in 1992. In 2002 the male and female cohorts both report unhealthier behavior than in 1992. Multivariate logistic regression analyses show that adjusted for age, cohort, lifestyle and psychosocial resources, poorly educated men have higher odds of functional limitations than well-educated men (OR = 2.62, 95% CI = 1.57-4.37). Analyses among women show a significant interaction effect between education and cohort. Poorly educated women have higher odds of functional limitations in 2002 than in 1992 (OR = 3.33, 95% CI = 1.02-10.87). Conclusions: The results underscore the need for policies focused on improving the health and lifestyle of the poorly educated. © Birkhäuser Verlag, Basel 2008.