Objectives. We examined the association between breastfeeding duration and maternal smoking before, during, and after pregnancy. Methods. Data from the 2000-2001 Oregon Pregnancy Risk Assessment Monitoring System were used. Early weaning was defined as not breastfeeding at 10 weeks postpartum. Results. At 10 weeks after pregnancy, 25.7% of mothers who initiated breastfeeding no longer breastfed. After controlling for confounders, quitters (mothers who quit smoking during pregnancy and maintained quit status after pregnancy) and postpartum relapsers (mothers who quit smoking during pregnancy and resumed smoking after delivery) did not have significantly higher risk for early weaning than nonsmokers. However, persistent smokers (mothers who smoked before, during, and after pregnancy) were 2.18 times more likely not to breastfeed at 10 weeks (95% confidence interval = 1.52, 2.97). Women who smoked 10 or more cigarettes per day postpartum (i.e., heavy postpartum relapsers and heavy persistent smokers) were 2.3-2.4 times more likely to wean their infants before 10 weeks than were nonsmokers. Conclusions. Maternal smoking is associated with early weaning. Stopping smoking during pregnancy and decreasing the number of cigarettes smoked postpartum may increase breastfeeding duration.