Aiming at the limitation of support vector machine (SVM) in the classification of large-scale intrusion signals,an improved SVM signal recognition method is proposed.Firstly,the particle swarm optimization algorithm (PSO) is used to generate the initial position with diversity.Then,the current position of the sample in the discrete search space is updated by the gray wolf optimization algorithm (GWO),and the optimal feature subset is obtained.Finally,the SVM is used to classify and recognize the samples to be tested based on the optimal feature subset.The experimental results show that the classifier based on PSO-GWO-SVM algorithm achieves the accuracy of 96.86%,sensitivity (SE) of 95.82% and specificity of 96.31%.Compared with the traditional signal recognition method,it has better recognition precision,adaptability and timeliness.