Background: Air pollution has been consistently linked with dementia and cognitive decline. However, it is unclear whether risk is accumulated through long-term exposure or whether there are sensitive/critical periods. A key barrier to clarifying this relationship is the dearth of historical air pollution data. Objective: To demonstrate the feasibility of modelling historical air pollution data and using them in epidemiological models. Methods: Using the EMEP4UK atmospheric chemistry transport model, we modelled historical fine particulate matter (PM2.5) concentrations for the years 1935, 1950, 1970, 1980, and 1990 and combined these with contemporary modelled data from 2001 to estimate life course exposure in 572 participants in the Lothian Birth Cohort 1936 with lifetime residential history recorded. Linear regression and latent growth models were constructed using cognitive ability (IQ) measured by the Moray House Test at the ages of 11, 70, 76, and 79 years to explore the effects of historical air pollution exposure. Covariates included sex, IQ at age 11 years, social class, and smoking. Results: Higher air pollution modelled for 1935 (when participants would have been in utero) was associated with worse change in IQ from age 11–70 years (β = −0.006, SE = 0.002, p = 0.03) but not cognitive trajectories from age 70–79 years (p > 0.05). There was no support for other critical/sensitive periods of exposure or an accumulation of risk (all p > 0.05). Conclusion: The life course paradigm is essential in understanding cognitive decline and this is the first study to examine life course air pollution exposure in relation to cognitive health.