English：

In subway tunnel construction, full⁃face tunneling machines are widely used for mechanizedexcavation. However, due to complex factors such as surface structures, settlement, and urban facilities,urban tunnel excavation faces unique challenges. Earth Pressure Balance （EPB） shield machines caneffectively reduce settlement risks, ensuring both excavation safety and improved efficiency. Amongthese, predicting the excavation rate is crucial for ensuring tunnel construction safety. The factorsinfluencing the EPB shield machine’ s excavation rate include geological, mechanical, and operationalparameters. Therefore, this study uses principal component analysis （PCA） to identify key factors, suchas torque, thrust, and speed,which explain over 95% of the variation in excavation rate. After removingoutliers and normalizing the data, excavation rate predictions are made using fuzzy logic based onMamdani and Sugeno algorithms, as well as a neuro⁃fuzzy method. The results show that the root meansquare error （RMSE） of the Sugeno algorithm is 23.51% lower than that of the Mamdani algorithm, andits performance index is approximately 6% higher. The neuro⁃fuzzy method shows a mean square error（MSE） that is 33. 14% and 12.58% lower than the Mamdani and Sugeno algorithms, respectively,demonstrating superior predictive performance.