BACKGROUND: Smoking during pregnancy has been suggested to have long-term consequences for neuropsychologic development in the offspring, including behavioral problems, attention deficit disorders, and antisocial behavior. Also, findings from several studies indicate an association with impaired cognitive function in the children. METHODS: In a population-based Swedish cohort study, we examined possible associations between maternal smoking in pregnancy and educational achievement in the offspring at age 15 years among more than 400,000 male and female students born 1983 through 1987. Generalized estimating equation models were used to evaluate associations of maternal smoking, other maternal characteristics, and birth characteristics with school performance. Odds ratios (ORs) were used as a measure of risk. RESULTS: In a model adjusted for maternal characteristics, maternal smoking compared with no tobacco use during pregnancy was associated with an increased risk of poor scholastic achievement: for 1-9 cigarettes per day, the OR was 1.59 (95% confidence interval = 1.59-1.63) and for 10 or more cigarettes per day, the OR was 1.92 (1.86-1.98). These risks remained unchanged when we also adjusted for smoking-related pregnancy outcomes such as fetal growth restriction and preterm birth. However, if the mother had smoked in her first pregnancy, but not in her second pregnancy, the younger sibling was also at increased risk of poor school performance. CONCLUSION: Observed associations between maternal smoking during pregnancy and poor cognitive performance in the offspring might not be causal. We suggest that associations reported in earlier studies may instead reflect the influence of unmeasured characteristics that differ between smokers and nonsmokers.