BACKGROUND AND OBJECTIVE: To analyze if cigarette smoking delays the sputum smear conversion in pulmonary tuberculosis. PATIENTS AND METHOD: Ninety eight patients were diagnosed with pulmonary tuberculosis. Patients were all not immunosuppressed, infected by human immunodeficiecy virus (HIV) or drug resistant. Sixty four of them were smokers with a pack-year index (standard deviation) of 33.69 (23.12). Delayed sputum smear conversion (DC) was considered when 2 positive sputum culture results were obtained in the second month of anti-tuberculous treatment and was associated with the following variables in 2 groups: a) total group (in which all the patients were included): age, sex, smoking habits, risk factors (alcohol consumption, diabetes mellitus, immunosuppression, drug addicion, malnutrition), time with symptoms, radiologic presentation and bacterial load, and b) smokers: age, sex, risk factors, time with symptoms, radiologic presentation, bacterial load and pack-year index. For the statistical analysis, chi(2) test, Student t test and logistic regression model were used, considering the dependant variable DC. RESULTS: In the total group, 17 patients (17.3%) had DC, 16 of them had a history of smoking and in the univariate analysis it was associated with: alcohol consumption, time with Symptoms, radiologic presentation as bilateral cavitary infiltrates and smoking habits. The logistic regression analysis showed an association with smoking habits (odds ratio = 9.8; p = 0.03) and bilateral cavitary infiltrates (odds ratio = 3.61; p = 0.02). In the group of smokers, DC was associated in the univariate analysis with the female sex. CONCLUSIONS: Smoking habits delay sputum conversion in patients with pulmonary tuberculosis not associated with HIV and non-resistant bacilli. According to these results it is necessary to assist smoking cessation in patients who are receiving antituberculous treatment.