Feature selection, as a preprocessing step to machine learning, is effective in reducing dimensionality, removing irrelevant data and increasing learning accuracy. The development of microarray dataset technology has supplied a large volume of data to many fields. In particular, it has been applied to prediction and diagnosis of cancer, so that it helps us to exactly predict and diagnose cancer. To precisely classify cancer we have to select genes related to cancer. The challenging task in cancer diagnosis is how to identify salient expression genes from thousands of genes in microarray data because extracted genes from microarray dataset have many unwanted datas not related to cancer. In this project we attempt explore a novelhybrid wrapper and filter feature selection algorithm for classification problem using a memetic framework i.e., a combination of genetic algorithm (GA) and local search (LS) has been proposed.. The LS is performed using correlation based filter methods are discritize, ranking and redundancy elimination with symmetrical uncertainty (SU) measure .using this hybrid method we can able find cancer related gene, From the larger amount of gene datas .using that smaller dataset doctors can able to find the affected gene and provide better treatment. The efficiency and the effectiveness of the method are demonstrated through extensive comparisons with other methods using real-world datasets of high dimentionality