The evolutionary approach for a decision of feature selection problem is considered in this paper. The new method of polymodal evolutionary search with a chromosome clusterization is offered. The developed method is based on idea of simultaneous search of several optimums, thus informative feature sets (chromosomes) are grouped in clusters on their arrangement in informative feature space that results in more uniform covering of feature search space. So stable subpopulations in different clusters are formed, diversity of search is provided, and convergence to different local minima is reached that allows to find more optimal combinations of features.