This study evaluated the long-term effects of different psychosocial intervention models in methadone maintenance treatment (MMT) in Xi'an China. Patients from five MMT clinics were divided into three groups receiving MMT only, MMT with counseling psychology (CP) or MMT with contingency management (CM). A five-year follow-up was carried out with daily records of medication, monthly random urine morphine tests, and tests for anti-HIV and anti-HCV every six months. Drug use behavior was recorded six months after initial recruitment using a survey. Adjusted RRs and their 95% confidence intervals (CIs) were estimated using an unconditional logistic regression model or a Cox proportional hazard model. A total of 2662 patients were recruited with 797 in MMT, 985 in MMT with CP, and 880 in MMT with CM. Following six months of treatment, the injection rates of MMT with CP and MMT with CM groups were significantly lower than that of MMT (5.1% and 6.9% vs. 16.3%, x² = 47.093 and 29.908, respectively; P<0.05). HIV incidences for MMT, MMT with CP and MMT with CM at the five year follow-up were 20.09, 0.00 and 10.02 per ten thousand person-years, respectively. HCV incidences were 18.35, 4.42 and 6.61 per hundred person-years, respectively, demonstrating that CP and CM were protective factors for HCV incidence (RR = 0.209 and 0.414, with range of 0.146-0.300 and 0.298-0.574, respectively). MMT supplemented with CP or CM can reduce heroin use and related risk behaviors, thereby reducing the incidence of HIV and HCV.