To evaluate the effectiveness of stent graft (SG) for the treatment of central venous disease (CVD) in hemodialysis patients. Between January 2014 and April 2022, 122 patients on hemodialysis with symptomatic CVD were treated with percutaneous transluminal angioplasty (PTA) and bare-metal stent (BMS) or SG placement. The mean follow-up was 14.5 months (IQR: 8.0, 24.2). Patency rates of the target sites were calculated using Kaplan-Meier and log-rank studies. Multivariate Cox proportional hazard models were used to evaluate the association between various characteristics and target site primary patency. Technical success rate was 100%. At 3, 6, 12, and 24 months, the target sites primary patency rates were 86.4%, 74.2%, 45.1%, and 30.4% for PTA; 94.7%, 78.6%, 60.8%, and 45.6% for BMS; and 94.0%, 92.0%, 82.4%, and 66.8% for SG, respectively, and the assisted primary patency rates were 86.5%, 80.4%, 63.8%, and 46.0% for PTA; 94.7%, 89.5%, 77.5%, and 71.1% for BMS; 100%, 100%, 97.8%, and 83.4% for SG, respectively. The Kaplan-Meier analysis indicated that SGs achieved better primary and assisted primary patency than PTA or BMS (p<0.05). SG use and concomitant stenosis were the independent predictors of target site primary patency dysfunction in the multivariate analysis. This study confirmed the better long-term patency of SG in comparison with PTA and BMS for the treatment of CVD in hemodialysis patients.