Background: Smoking is a risk factor for mucinous ovarian cancer (OvCa) in Caucasians. Whether a similar association exists in African Americans (AA) is unknown. Methods: We conducted a population-based case–control study of incident OvCa in AA women across 11 geographic locations in the US. A structured telephone interview asked about smoking, demographic, health, and lifestyle factors. Odds ratios and 95% confidence intervals (OR, 95% CI) were estimated from 613 cases and 752 controls using unconditional logistic regression in multivariable adjusted models. Results: Associations were greater in magnitude for serous OvCa than for all OvCa combined. Compared to never smokers, increased risk for serous OvCa was observed for lifetime ever smokers (1.46, 1.11–1.92), former smokers who quit within 0–2 years of diagnosis (5.48, 3.04–9.86), and for total pack-years smoked among lifetime ever smokers (0–5 pack-years: 1.79, 1.23–2.59; >5–20 pack-years: 1.52, 1.05–2.18; >20 pack-years: 0.98, 0.61–1.56); however, we observed no dose–response relationship with increasing duration or consumption and no significant associations among current smokers. Smoking was not significantly associated with mucinous OvCa. Associations for all OvCa combined were consistently elevated among former smokers. The proportion of ever smokers who quit within 0–2 years was greater among cases (23%) than controls (7%). Conclusions: Cigarette smoking may be associated with serous OvCa among AA, which differs from associations reported among Caucasians. Exposure misclassification or reverse causality may partially explain the absence of increased risk among current smokers and lack of dose–response associations. Better characterization of smoking patterns is needed in this understudied population.