Introduction: The nature of the relationship between maternal tobacco smoking during pregnancy and the occurrence of children’s behavioural problems is still a matter of controversy. We tested this association using data collected among a sample of pregnant women and their offspring followed from birth to early adolescence (age 12 years), accounting for multiple parent, child, and family characteristics. Methods : Data come from 1424 mother-child pairs participating in the EDEN mother-child cohort in France. Using repeated measures (3, 5.5, 8 and 11.5 years) of the mother-reported Strengths and Difficulties Questionnaire, we estimated trajectories of children’s emotional and behavioural difficulties. Two aspects of maternal smoking were studied: the timing (non-smoker, smoking during the periconceptional period or throughout pregnancy) and the level of use (cigarettes/day) during the first trimester of pregnancy. Robust Poisson regression models controlled for confounding factors including maternal mental health and socioeconomic characteristics using propensity scores with the overlap weighting technique. Results: Contrary to bivariate analyses, in propensity score-controlled regression models, maternal smoking throughout pregnancy was no longer significantly associated with offspring emotional or behavioural difficulties. Maternal heavy smoking (≥10cigarettes/day) remained significantly associated with intermediate levels of conduct problems (RR 1.25 95%CI 1.19-1.31) Conclusion: The association between maternal smoking in pregnancy and offspring emotional and behavioural difficulties appears to be largely explained by women’s other characteristics. However, maternal heavy smoking appears to be related to offspring behavioural difficulties beyond the role of confounding characteristics.