Abstract: To improve the accuracy and real-time performance of intrusion detection models in complex net-work environments, a network intrusion detection model based on random forest and eXtreme Gra-dient Boosting (XGBoost) is proposed. First, the feature importance is calculated based on the ran-dom forest algorithm. A hybrid feature selection method combining filtering and embedded is used to reduce the feature dimension of the dataset. When detecting the sample category, the XGBoost algorithm based on cost-sensitive function and grid method tuning is used to improve Model accu-racy. Experimental simulation results show that compared with other machine learning algorithms, the proposed model greatly reduces processing time by more than 50% with higher detection accu-racy, and has better robustness and adaptability.