Evidence about the impact of air pollution on cognitive development of children has been growing but remains inconclusive. To investigate the association of air pollution exposure and the cognitive development of children in the UK Millennium Cohort Study. Longitudinal study of a nationally representative sample of 13 058–14 614 singleton births, 2000–2002, analysed at age 3, 5 and 7 years for associations between exposure from birth to selected air pollutants and cognitive scores for: School Readiness, Naming Vocabulary (age 3 and 5), Picture Similarity, Pattern Construction (age 5 and 7), Number Skills and Word Reading. Multivariable regression models took account of design stratum, clustering and sampling and attrition weights with adjustment for major risk factors, including age, gender, ethnicity, region, household income, parents’ education, language, siblings and second-hand tobacco smoke. In fully adjusted models, no associations were observed between pollutant exposures and cognitive scores at age 3. At age 5, particulate matter (PM2.5, PM10), nitrogen dioxide (NO2), sulphur dioxide (SO2) and carbon monoxide (CO) were associated with lower scores for Naming Vocabulary but no other outcome except for SO2 and Picture Similarity. At age 7, PM2.5, PM10 and NO2 were associated with lower scores for Pattern Construction, SO2 with lower Number Skills and SO2 and ozone with poorer Word Reading scores, but PM2.5, PM10 and NO2 were associated with higher Word Reading scores. Adverse effects of air pollutants represented a deficit of up to around four percentile points in Naming Vocabulary at age 5 for an interquartile range increase in pollutant concentration, which is smaller than the impact of various social determinants of cognitive development. In a study of multiple pollutants and outcomes, we found mixed evidence from this UK-wide cohort study for association between lifetime exposure to air pollutants and cognitive development to age 7 years.