Objective. This research aims to determine whether residence (rural vs. urban) at different life stages (childhood, adulthood, and late life) is associated with increased risk of incident dementia in a population-based cohort of older Spaniards. Methods. In this prospective study, 2,711 participants aged 65 years and older were assessed at baseline and 3 years later. All cases of incident dementia were diagnosed using DSM-IV criteria. The relationship between residence and the relative risk of dementia was analysed using Cox's regression models. Demographics, comorbidity index, consumption (tobacco / alcohol) and doubtful dementia diagnosis were considered as possible confounders. Results. At the three-year follow-up, 91 cases of dementia were detected. Lower education and occupational attainment were associated with a higher incidence of dementia three years later. Rural residence in adulthood was associated with a significantly higher risk of dementia at the follow-up. Childhood rural residence revealed a marked trend for risk of dementia (p = 0.08), but it was nonsignificant in later life. The risk of dementia was considerably higher for the rural/low-education group than for the urban/high-education group, for both childhood and adulthood residence. Finally, people from areas with the lowest socio- economic status Arévalo (rural, blue-collar) and Margaritas (urban, blue-collar) showed higher risk of dementia than people from Lista (urban, mixed white/blue collar). Conclusion. In this cohort, early and mid-life stages rural residence was a risk factor for dementia, but not later-life rural residence. The rural residence effect was noticeably higher in people with a lower educational level.