In recent years, as the usage of internet increases, new types of attack on network information keep increasing. Intrusion Detection System (IDS) is an important tool to identify various attacks to secure the networks. Traditional clustering algorithms work on “flat” data, making the assumption that the data instances are homogeneous in nature. Many real world data, however, is heterogeneous in nature. To handle the heterogeneity of the data, in this paper, we propose IDS using K-SVMeans clustering algorithm and other classification techniques. Here we use KDD-Cup99 dataset as simulation dataset for our experiment. The experimental results show that the proposed model reduces the time complexity, improves overall detection accuracy and minimizes the false alarm rate considerably.