Abstract: Various analytical and empirical formulas and models have been developed for predicting tunnel water inflow; however, discrepancies often exist between calculated results and actual measurements.To enhance accuracy of tunnel water inflow prediction, this study employs the entropy method to calculate weights and integrates these with the ordered weighted averaging method for combined prediction of individual model outputs.Based on the actual monitoring data of Liebaituo Tunnel in Yunnan Province, five commonly used single prediction models and two combined prediction models were evaluated based on effectiveness and other performance indicators.The results demonstrated that the entropy-based combined prediction method achieved an average accuracy of 96.46% across five prediction segments, with an effectiveness value of 0.941 9.Compared with the most accurate single prediction model, this combined approach improved average accuracy by 18.17% and effectiveness by 24.6%.The combined prediction method can significantly improve prediction accuracy, provide support for formulating tunnel drainage scheme, and reduce the disaster risk caused by tunnel water inflow.