INTRODUCTION: In 2012, Bangladesh continues to be one of the 22 high tuberculosis (TB) burden countries in the world. Although free diagnosis and management for TB is available throughout the country, case notification rate/100,000 population for new smear positive (NSP) cases under the national TB control programme (NTP) remained at around 70/100,000 population and have not changed much since 2006. Knowledge on TB disease, treatment and its management could be an important predictor for utilization of TB services and influence case detection under the NTP. Our objective is to describe knowledge of TB among newly diagnosed TB cases and community controls to assess factors associated with poor knowledge in order to identify programmatic implications for control measures. METHODS: Embedded in TB prevalence survey 2007-2009, we included 240 TB cases from the TB registers and 240 persons ≥ 15 years of age randomly selected from the households where the survey was implemented. All participants were interviewed using a structured, pre-tested questionnaire to evaluate their TB knowledge. Regression analyses were done to assess associations with poor knowledge of TB. RESULTS: Our survey documented that overall there was fair knowledge in all domains investigated. However, based on the number of correct answers to the questionnaires, community controls showed significantly poorer knowledge than the TB cases in the domains of TB transmission (80% vs. 88%), mode of transmission (67% vs. 82%), knowing ≥ 1 suggestive symptoms including cough (78% vs. 89%), curability of TB (90% vs. 98%) and availability of free treatment (75% vs. 95%). Community controls were more likely to have poor knowledge of TB issues compared to the TB cases even after controlling for other factors such as education and occupation in a multivariate model (OR 3.46, 95% CI: 2.00-6.09). CONCLUSIONS: Knowledge on various aspects of TB and TB services varies significantly between TB cases and community controls in Bangladesh. The overall higher levels of knowledge in TB cases could identify them as peer educators in ongoing communication approaches to improve care seeking behavior of the TB suspects in the community and hence case detection.