This study proposes an extension of k-tuple method which utilizes the ratio of frequency of common sub words of length k to compare two sequences. The proposed method has two stages. stage 1 extracts feature from the sequence to obtain distance matrix and stage 2 obtains clusters from similarity matrix. The proposed method is tested on four datasets and the results are compared with those of k-tuple and tree generated using clustalw. Purity of tree and symmetric distance between the tree generated from proposed method and alignment based methods have also been computed. The results of proposed method are also compared with Composition Vector and k-tuple.