Abstract： Objective:To construct a non-invasive predictive model for liver fibrosis in HBeAg positive chronic hepatitis B (CHB) patients with alanine aminotransferase (ALT) lower than 2 upper limit of normal (ULN).Methods:The clinical data of 279 HBeAg positive CHB patients with ALT<2×ULN admitted in Zhejiang Provincial People’s Hospital from October 2014 to December 2020 were retrospectively analyzed. According to the pathological results of liver biopsy, there were 117 cases of mild liver fibrosis (S0-S1) and 162 cases of significant liver fibrosis (S2-S4). The independent predictors of liver fibrosis were analyzed by multivariate logistic regression analysis and a noninvasive predictive model was constructed. The model for predicting the severity of liver fibrosis was evaluated by receiver operating characteristic curve (ROC).Results:Multivariate logistic regression analysis showed that age, prothrombin time (PT), aspartate aminotransferase (AST), anti-HBc and HBV DNA were independent predictors of liver fibrosis ( OR=1.055, 1.365, 1.027, 1.231, 0.763, all P<0.05). The area under the ROC curve (AUR) of the model was 0.772 (95% CI: 0.716-0.828, P<0.05), the sensitivity and specificity for the diagnosis of significant liver fibrosis were 79.5% and 70.9% at the cut-off value of 0.504. The AUC of APRI model and FIB-4 index model for assessing significant liver fibrosis in CHB patients with HBeAg-positive and ALT<2×ULN were 0.720 and 0.671, respectively, which were lower than that of the current model (all P<0.05). Conclusion:The noninvasive predictive model constructed in this study has a high diagnostic value for evaluating the severity of liver fibrosis in CHB patients with HBeAg positive and ALT<2×ULN.