Pattern classification is a very powerful clustering and classification of algorithms consisting of an efficient classifier to classify the genomic data set with high accuracy, time and memory complexity of genomic domain The significance of genome sequence clustering lies on the basics of molecular biology, genome sequence alignment and similarity scores. For sequence dataset, motif string/sequences can be used as a class/ cluster representative. Unsupervised local alignment algorithm is proposed for genome sequences classification and it performs better with global alignment based approaches. Performance of unsupervised motif based clustering algorithm is also evaluated for genome sequences dataset and it also performs well, but the time and space complexities of genome sequences alignment algorithm are quadratic. A simple feature selection technique for reducing the time and space requirements for genome sequence comparison is proposed. It can be used for whole clustering or classifying the genome sequences based on sequence similarity without much reduction in the CA. Incremental clustering for large dataset, analysis of algorithm and properties related to leader based