STUDY OBJECTIVE--The aim was to study changes in smoking habits during pregnancy and differences in characteristics between women who stop smoking and those who continue to smoke during pregnancy. DESIGN--The study was a population based prospective study. Self administered questionnaires were completed on three occasions. SETTING--The study area was Uppsala county, Sweden, in 1987. PARTICIPANTS--The participants were women registered with antenatal care clinics, which included all pregnant women in the county. Ninety six percent (n = 3678) of all pregnant women completed the first questionnaire. Thirty two percent of these were smokers at time of conception. MEASUREMENTS AND MAIN RESULTS--Twenty nine percent of the smokers stopped smoking at some stage of pregnancy, and the majority did so before having registered for antenatal care. Using logistic regression analysis it was found that high parity number, not living with infant's father, heavy smoking, and daily passive smoking at home were associated with significantly increased risk for continued smoking during pregnancy. High level of education and high age at onset of smoking decreased the risk. CONCLUSIONS--In order to reduce the smoking related risks for unsuccessful pregnancy outcome, general preventive efforts in society must be combined with the development of more specialised antenatal programmes designed with consideration of the characteristics and life situation of the pregnant smoker.