Abstract: Genetic Algorithm(GA) is used to optimize the selected elements from a full array,so as to achieve narrow mainlobe and minimized sidelobes pattern.Setting the working sate and excitation of elements as variables,mainlobe width and sidelobe level as fitness function of GA,fitness ranking and multi-populations competition techniques were applied to assure that the GA converge to the optimum result.The effects of layouts,sparse rate and the excitation to the pattern of thinned array are analyzed in detail.