The peak particle velocity(PPV) of blasting vibration is an important index to measure the impact of blasting vibration on surrounding environment and structures.In order to improve the reliability of PPV prediction,a model based on extreme gradient boosting optimized by the sparrow search algorithm was proposed,and a corresponding blasting vibration prediction system was built using the App Designer of MATLAB.The maximum charge per delay,distance from blast center to measuring point,and elevation difference between measuring point and blast center were selected as the input parameters of 36 sets of training data and 5 sets of test data for the model to predict PPV.The results show that the proposed SSA-XGBoost model has a smaller average relative error compared with the GA-BPNN model and BPNN model,and it has a higher prediction accuracy and better stability proved by the Taylor graph.