A multi parameter coupled monitoring model that integrates temperature, kinematic viscosity, micro water content, electrical resistivity, water activity, and pollution level to address the engineering challenge of abnormal increase in total acid value (TAN)  caused by hydrolysis degradation in the EH fire-resistant oil system of power plant steam turbines were constructed. By analyzing the hydrolysis kinetics of EH oil and establishing a temperature compensation mechanism based on the arrhenius equation, combined with multiple linear regression and stepwise variable selection algorithm, dynamic prediction of acid value was achieved. The experimental results show that under typical operating conditions of 25、40 and 100 ℃, the root mean square error (RMSE) predicted by the model is 0.039, 0.055, and 0.068 mgKOH/g, respectively, and the coefficient of determination (R2) exceeds 0.93, which is significantly better than traditional offline detection and single parameter monitoring methods. Relying on the online monitoring system developed by the distributed sensor network and edge computing platform, 12 months of continuous monitoring was achieved in a 600 MW unit, successfully extending the oil replacement period by 50%, and the early warning accuracy rate reached 92%. The research results provide a new technological path for the accurate evaluation of EH anti fuel deterioration status, which has important engineering significance for improving the reliability of the turbine electro-hydraulic control system.