Objective: The purpose of this study was to explore the relationships between perceived social support and retention in Chinese mainland patients receiving methadone maintenance treatment(MMT).Methods: This was a cross-sectional two-year follow-up study. The data collected included patients’ baseline characteristics, perceived social support and retention in MMT.Results: A total of 1212 patients completed the cross-sectional survey; 809(66.7%) had good perceived social support and 458(37.8%) had experienced readmissions. With and without controlling for baseline characteristics, past retention had no significant influence on perceived social support. By the end of the follow-up, 527(43.5%) patients had terminated MMT. The patients without good perceived social support were more likely to terminate treatment than those with good perceived social support [hazard ratio:1.31, 95% confidence interval: 1.10, 1.57; 1.25(1.04, 1.51)] regardless of their baseline characteristics and past retention.Conclusions: Retention, thus, had no significant influence on perceived social support in MMT, whereas good perceived social support was a strong protective predictor of retention. Objective: The purpose of this study was to explore the relationships between perceived social support and retention in Chinese mainland patients receiving methadone maintenance treatment (MMT). Methods: This was a cross-sectional two-year follow-up study. patients’ baseline characteristics, perceived social support and retention in MMT. Results: A total of 1212 patients completed the cross-sectional survey; 809 (66.7%) had good perceived social support and 458 (37.8%) had experienced readmissions. With and without controlling for baseline characteristics, past retention had no significant influence on perceived social support. By the end of the follow-up, 527 (43.5%) patients had terminated MMT. The patients without good perceived social support were more likely to terminate treatment than with good perceived social support [hazard ratio: 1.31, 95% confidence interval: 1.10, 1.57; 1.25 (1.04, 1.51)] regardless of their baseline characteristics and past retent ion.Conclusions: Retention, thus, had no significant influence on perceived social support in MMT, respect good perceived social support was a strong protective predictor of retention.