OBJECTIVE Maternal postpartum smoking is associated with several negative outcomes for mother and child. The relationship between bonding impairment (BI) and maternal postpartum smoking is less known. The aim of the present study was to assess the relationship between BI and maternal postpartum smoking at 6 to 8 months after childbirth in a primary care sample of low-income mothers who had depression during the previous pregnancy. METHODS We performed a cross-sectional, secondary analysis from a community trial with 356 postpartum women from public primary care clinics in São Paulo, Brazil. The main outcome measure was maternal postpartum smoking. The main exposure variable was BI assessed with the Postpartum Bonding Questionnaire. A socio-demographic questionnaire evaluated maternal and obstetric characteristics. Postpartum depression was assessed with the Patient Health Questionnaire. Crude and adjusted prevalence ratios (PR) of the association between BI and maternal postpartum smoking, with 95% confidence intervals (CI), were calculated using Poisson regression. Multivariate analysis was performed using three models (Model 1 adjusted for randomization during pregnancy, Model 2 adjusted for Model 1 plus mother's socioeconomic and obstetric characteristics, and Model 3 adjusted for Model 2 plus postpartum depression). RESULTS Maternal postpartum smoking occurred in 16.7% of our sample. In the multivariable analysis, BI was associated with maternal postpartum smoking (PR:2.04; CI 95% 1.08:3.84). CONCLUSIONS Women presenting bonding problems are at higher risk of smoking after childbirth. Intervention to enhance mother-child bonding may have the potential to decrease maternal postpartum smoking.