One of the principal objectives of tuberculosis (TB) control is to minimise the emergence of drug resistance. The first national survey was conducted in Rwanda to determine the prevalence of M. tuberculosis drug resistance. Sputum samples were collected from all new and retreatment cases in the health districts from November 2004 to February 2005. Drug susceptibility testing of isolates against first-line drugs was performed by the proportion method. Of 616 strains from new cases, 6.2% were resistant to isoniazid, 3.9% to rifampicin and 3.9% were multidrug-resistant TB. Among 85 strains from previously treated cases, the prevalence of resistance was respectively 10.6%, 10.6% and 9.4% (MDR-TB strains). Eight MDR cases showed additional resistance to ethambutol and streptomycin. The level of MDR-TB among TB patients in Rwanda is high. The main reasons of this emergence of MDR-TB can be attributed to the disorganisation of the health system, migration of the population during the 1994 civil war and poor success rates, with a high number of patients transferred out and lost to follow-up. On the other hand, the use of treatment regimens administered twice weekly during the continuation phase could be another important factor and merit further investigations.