Objective: To assess the overall mortality of methadone maintenance treatment (MMT) clients in China and its associated factors. Methods: A total of 1,511 MMT clients, all of whom enrolled in China's first eight MMT clinics between March and December 2004, were included in this cohort study and followed for approximately six years, until June 2010. Mortality and its predictors were examined using Cox proportional hazards regression models. Results: A total of 154 deaths were observed within 5,391 person-years (PY) of follow-up for an all-cause mortality rate of 28.6 per 1,000 PY. The leading causes of death were drug overdose (33.8%), HIV/AIDS-unrelated disease (21.4%), and HIV/AIDS (16.9%). The all-cause mortality rate of clients engaged in MMT for one year or less was roughly three times that of clients who stayed in MMT for four years or more (14.0 vs. 4.6, p<0.0001), HIV-positive subjects was nearly four times mortality rate than that of HIV-negative individuals (28.1 vs.6.8, p<0.0001). ART-naive HIV-positive subjects had approximately two times higher mortality rate than those receiving ART (31.2 vs. 17.3, <0.0001). After adjusting for confounding variables, we found that being male (HR = 1.63, CI: 1.03-2.57, p = 0.0355) and being HIV-positive (HR = 5.16, CI: 3.70-7.10, p<0.0001) were both associated with higher risk of death whereas increased durations of methadone treatment were associated with a lower risk of death (HR = 0.26, CI: 0.18-0.38, p<0.0001 for two to three years, HR = 0.08, CI: 0.05-0.14, p<0.0001 for four or more years). Conclusion: Overall mortality was high among MMT clients in China. Specific interventions aimed at decreasing mortality among MMT clients are needed. Our study supports the need for keeping client at MMT longer and for expanding ART coverage and suggests the potential benefits of integrated MMT and ART services for drug users in China. © 2013 Cao et al.