Abstract: A new clustering algorithm and genetic algorithm based on a combination of intrusion detection model.Algorithm to cluster the center of the binary code of each point to their respective cluster centers between the Euclidean distance as similarity measure integrated, and then use genetic algorithm, effectively reduce the path length of the network topology, the genetic algorithm to find the cluster center.Matlab simulation results show that the improved network anomaly detection, and the more traditional model of network intrusion detection system, compared with the invasion of a better recognition rate and detection rate, while increasing the efficiency of the algorithm.