Intrusion detection systems are intelligent systems designed to identify and prevent the misuse of computer networks and systems. Various approaches to Intrusion Detection are currently being used, but they are relatively ineffective. Thus the emerging network security systems need be part of the life system and this is possible only by embedding knowledge into the network. The Adaptive Genetic Algorithm Model - IDS comprising of K-Means clustering Algorithm, Genetic Algorithm and Neural Network techniques. The technique is tested using multitude of background knowledge sets in DARPA network traffic datasets.