Abstract: To solve the problem of intrusion detection false positives and intrusion detection accuracy caused by the rapid increase in the number and diversity of smart devices connected to the Internet in open WSN, an intelligent lightweight intrusion detection algorithm based on enhanced support vector machine (ESVM) for classification and genetic algorithm (GA) for feature selection is proposed. The algorithm transformed the complex traffic of intrusion dataset into the readable format of SVM, and used crossover and mutation operators to intelligently select the traffic characteristics with the largest amount of information to reduce the dimension of wireless network traffic. It used ESVM to perform classification to identify intrusion detection more effectively. The implementation results show that the algorithm has obvious improvement in selecting the optimal flow and improving the detection accuracy.