To enhance the accuracy and stability of fault prediction for mining excavator engines under high-intensity and complex working conditions, a fault prediction model based on the Grey Wolf Optimizer (GWO) optimized Support Vector Machine (SVM) was constructed. Key operating parameters of the Weichai WP14T mining excavator engine were collected during real-time operation. Data preprocessing was conducted using outlier detection and elimination as well as data standardization. Feature selection was performed using the Pearson correlation coefficient method to construct a sample set with a reasonable structure and complete feature information. The Grey Wolf Optimizer was employed to adaptively and globally optimize the penalty factor and kernel function parameters of the Support Vector Machine, thereby enhancing the model's nonlinear learning and generalization capabilities. Fault prediction experiments were designed, and comparative verification was carried out under various typical fault conditions. The results show that the GWO-SVM model exhibits higher accuracy and stronger robustness in multi-class fault prediction. The prediction accuracy on the test set reaches 92.8%, which is 9.9 percentage points higher than that of the traditional SVM, and it maintains stable output under conditions of few-sample interference and data fluctuation.