A weighted real-coded genetic algorithm is proposed in this paper, in allusion to the characteristics of optimizing designs of metamaterials, which have many parameters and parameters of different powers in metamaterial structure. The algorithm utilizing allele or double gene to change the power of one gene, develops the idea of weighted encoded binary coding. Compared with common real-coded based genetic algorithm, the weighted real-coded genetic algorithm adds artificial selection in it, it can not only accelerate the speed of convergence, but also improve the solution quality, especially for the design of large-scale and long computation time. In this paper, the algorithm is verified by optimizing a metamaterial absorber.