Objectives Bedaquiline treatment significantly improves multidrug-resistant tuberculosis (MDR-TB) patient treatment outcomes. However, safety and efficacy data are lacking for bedaquiline used with background regimens to treat Chinese TB patients. Here, we describe our initial clinical experience for bedaquiline treatment of a large multicentre cohort of MDR-TB and extensively drug-resistant tuberculosis (XDR-TB) patients in China. Methods Patients (177) received 24-week bedaquiline treatment combined with personalized anti-TB drug background regimens. As primary efficacy endpoints, times to initial sputum culture conversion were measured. Results Of 177 MDR-TB patients completing the 24-week treatment course, sputum culture conversion occurred for 151/177 (85.3%), while 26 had unfavourable outcomes, including 3/177 (1.7%) deaths and 23/177 (13.0%) non-responders at treatment completion. The median time to sputum culture conversion was 4 (interquartile range 2–8) weeks. Conversion rates were 33/39 (84.6%, 95% confidence interval (CI) 73.3–95.9) for MDR-TB patients, 47/56 (83.9%, 95% CI 74.3–93.6) for pre-XDR-TB patients and 71/82 (86.6%, 95% CI 79.2–94.0) for XDR-TB patients. Multivariate analysis demonstrated that patients with low body mass index (odds ratio 7.356; 95% CI 2.652–20.401) were at significantly high risk of unfavourable outcomes, with serious adverse events noted in 15 (8.5%) patients, including six with corrected QT interval (QTc) prolongation times (>500 ms). Conclusion Bedaquiline, when included in background regimens for treatment of MDR-TB and XDR-TB patients in China, was safe and associated with a high rate of culture conversion.