100 Abstract: Security of Wireless Ad hoc network has a primary concern to provide protected communication between mobile nodes. When we routing some packet it can use both malicious node or authenticate node for forwarding and receiving data. Malicious node can attack like black hole, misuse of data or hacked information. Our aim is to discuss the feasibility of monitoring the node of different networks, to analyze it for providing better security in AODV routing protocol. We implement data mining techniques for search large amount of data according characteristic rules and patterns to detect malicious node. We have used growing neural gas (GNS) clustering algorithm to make clusters and analysis data. Using soft computing technique we find patterns, analysis node and take decision based on decision tree.