Purpose. To determine the prevalence of major depressive disorder (MDD) by smoking status, and the temporal relationship between smoking and MDD, and explore other smoking-related variables that may be associated with MDD. Design. Cross-sectional study. Setting. Public primary care clinics. Subjects. Researchers studied 338 women (76% African-Americans) who were randomly selected while attending appointments in two public primary care clinics. Measures. Data pertaining to smoking-related variables and MDD diagnosis were obtained using the Diagnostic Interview Schedule for the Diagnostic Statistical Manual of Mental Disorders IV (DSM-IV). Results. The prevalence of a lifetime history of MDD was significantly higher for current smokers (56.6%) than among former smokers (37.5%) or never-smokers (30.3%; p < .001). Most ever-smokers (81.3%) began smoking and were nicotine-dependent (63.6%) prior to their first episode of MDD. Using logistic regression, after controlling for demographic and smoking-related variables, age of smoking onset was the strongest variable associated with MDD among ever-smokers. Specifically, the odds of having an MDD decreased by 8.2% for each year delay, in smoking initiation. Conclusion. These results suggest that smoking initiation precedes MDD and that smoking is associated with a high prevalence of MDD among low-income women attending primary care clinics. Further, the younger women start smoking the more likely they are to have MDD.