In recent decades, wireless sensor networks have faced problems. One of the main problems is security. Subsequent advances in network connectivity and computing capabilities have expanded the applications of wireless sensor networks (WSNs). Data collection and transmission to a remote server, located in isolated locations, is the main goal of wireless sensor networks (WSNs). These networks may be underground, underwater, terrestrial or multimodal. Among them are industrial automation, traffic management, monitoring of medical devices and other areas. Despite the growth of the market, wireless sensor networks are facing challenges. Energy efficiency, manufacturing and processing boundaries, bandwidth, error rate, stored resources and survival under harsh conditions should be considered. Using some network security methods, it is possible to deal with the target, but these methods have a negative effect on the lifetime of network sensors and cause a loss of overall network energy. In this article, by simulating a wireless sensor network, we first use the k-means clustering algorithm to divide the nodes into different clusters in terms of their distance from the other nodes and cluster heads, as well as the energy consumption of each node for Sending messages and the distance from the base station in the most optimal possible state. Then, by simulating the Smurf attack, several nodes send too many messages, which causes the energy consumption of the node, then by creating an intrusion detection system, the attacking nodes are identified and removed. The innovation of this article is that in addition to identifying other attacking nodes whose status is not normal, it is identified and considered as suspicious nodes, and in the end, it increases the overall lifespan of the collection by removing the attacking nodes.