The problem of computer network security is the very hard of detecting new attacks which do not have known signatures of intrusion. Intrusion Detection Systems (IDS) is a program of monitoring the events in a computer network and analyzing them for signature of intrusions. This paper proposed the clustering technique by using hybrid method based on Principal Component Analysis (PCA) and Fuzzy Adaptive Resonance Theory (FART) for identifying various attacks. The PCA is applied to random selects the best attribution and reduction the feature space. FART is implementing used to classifying difference group of data, Normal and Anomalous. The results show that the proposed technique can improve the high performance of the detection rate and to minimize the false alarm rate. The evaluated our approach on the benchmark data from KDDCup'99 data set.