Abstract: Objective To establish a model for predicting the short-term prognosis of patients with HBV-related acute-onchronic liver failure (HBV-ACLF) based on red blood cell distribution width (RDW) and the model for end-stage liver disease (MELD) scores. Methods A total of 245 patients with HBV-ACLF were retrospectively analyzed for their clinical data and results of routine hematological tests, liver function, renal function, coagulation test, HBV-DNA, and other indicators at admission. Univariate analysis and binary logistic regression analysis were used to test the short-term risk factors for death of the patients, and the MELD-RDW model was established. The accuracy of each index and the established model was verified using the ROC curve. Results The surviving patients with HBV-ACLF had significantly decreased RDW (14.97 ± 1.38) and MELD score (23.54±4.35) compared with those in the patients dead within 90 days (17.05±2.92 and 28.95±5.99, respectively). Multivariate analysis indicated that RDW was a significant independent prognostic factor for mortality in patients with HBVACLF (OR=1.840, 95%CI: 1.47902.289, P<0.005). The risk assessment model was [logisticMELD-RDW] =- 9.375 + 0.582 ×RDW- 0.091×ALB-0.05×PTA+0.186×MELD. The area under the ROC curve of MELD score combined with RDW was 0.878, which was higher than RDW (0.724) and MELD score (0.780) alone. Conclusion RDW is an independent prognostic indicator for mortality in patients with HBV-ACLF. Compared with MELD score, the risk assessment model based on MELD and RDW has a greater value in predicting the short-term prognosis of patients with HBV-ACLF.