Objectives: To identify clinical, radiological and microbiological factors associated with the non conversion of sputum smears in new smear positive cases of pulmonary tuberculosis after two months of treatment and to evaluate the influence of non-smear conversion on treatment outcomes.Design: A prospective cohort study.Setting: Tuberculosis centre of Hôpital Jamot in Yaoundé- Cameroon from April 2006 to September 2007.Subjects: A total of 413 patients were studied.Main outcome measures: Sputum smear status at two months of treatment, favourable treatment outcome (cured, treatment completed), unfavourable treatment outcome (death, treatment failure, default from treatment) and transferred out.Results: A total of 413 patients were studied; There were 234 (56.8%) males and 178 (43.2%) females with a mean age of 33 years (range 9.80 years). Sputum smears did not convert in 55 (13.4%) patients at the end of two months of treatment. Logistic regression analysis showed that age above or equal to 40 years (OR=2.716, 95% CI:1.412- 5.223, p = 0.003), and a bacillary load of 3+ on pre-treatment sputum smears (OR=1.955; 95%CI: 1.039-3.68, p=0.037) were significantly associated with non conversion of sputum smears at the end of two months of treatment. Persistent positive smears at the end of two months of treatment were significantly associated with unfavourable treatment outcomes (p=0.025) especially default during the course of treatment.Conclusion: In Yaoundé, Cameroon, non conversion of positive sputum smears in new patients with pulmonary tuberculosis at the end of two months of treatment is associated with an unfavourable outcome particularly defaulting later in the course of treatment. Non conversion of sputum smears at two months of treatment is significantly associated with age above or equal to 40 years and the presence of numerous bacilli (3+) on pre-treatment sputum smears. Patients with these factors who do not smear convert after two months of treatment should be given a fully supervised treatment for the entire duration of therapy so as to prevent in particular treatment default.