Due to recent changes of computer & networks, the IDS(Intrusion Detection System) need to be developed for new intrusion patterns. The current IDS have limited on recognition and correspond to new intrusion patterns on detection speed for multi packet which dealing on the network. Therefore, new technology need to increasing efficiency and speed of detection speed requested. The aim of this research is the development of standard and systematic method on intrusion detection. The core idea is using data mining method to find bundle of patterns on networking program and user behavior patterns as well as apply the feature systems to calculate the classifiers which could recognize the well known or irregular intrusions. In this paper, we will recommend following steps to develop the intrusion detection system: First, we will learn the detection applying technique for multi intrusion cases. Second, we will use data mining technique which fast recognize the current intrusion patterns and new patterns. Third, to recognize intrusion patterns, information of packet on the network and recorded data on the host sessions have studied. Fourth, we will create regulations between intrusion and normal behavior by practical use of logged file abstraction programs. Fifth, we will analysis intrusion detection pattern based on the created regulations and study results.