We used a nested case-control design with study participants sampled from two cohorts, for a total of 2,198 elderly people 65 years or older and completed cognitive tests between 1993 and 1997, to assess the association between an individual's lifetime principal occupation and the subsequent risk of cognitive impairment. Cases consisted of 290 older adults with impaired cognitive functioning. For each case, two controls with comparable age (within 5 years) and sex frequencies were randomly sampled from the seniors free of cognitive impairment. Occupational data were collected through interviews. Individual's job content was coded into one of the occupational categories or the occupation-based social classes. Compared to those who were former legislators, government administrators, or business executives and managers, a significantly elevated risk of cognitive impairment was estimated for those who were employed as agriculture/animal husbandry/forestry/fishing workers (odds ratio (OR)=3.2), craft and related trades workers (OR=2.2), plant and machine operators and assemblers (OR=14.7), workers of elementary occupations (OR=3.2), or housekeepers (OR=2.6). We also observed health inequalities in the risk of cognitive impairment across social classes with a significant dose-response trend in which unskilled blue-collar workers had the highest risk. After adjustment for education, we still observed an inverse relationship between risk of cognitive impairment and occupational class. This may mean that lifetime longest-held occupation is more intimately involved in the causal pathways leading to cognitive impairment. Further studies that collect information on specific work hazards would help make specific interpretations of the observed effect of lifetime longest-held occupation in early adulthood on risk of cognitive decline in late life.