The optimization design model of embedded giant magnetostrictive components (EGMC) was presented with multi-field property of machinery, electric, magnetic and thermal. And the multi-object genetic algorithm was applied to the optimization design of the EGMC. All optimal objects of the model were confirmed by the general design rules and the requirements of non-cylindrical holes in precision processing. The optimal goals of the model were as follows: a suitable flexural rigid of the EGMC, the maximum torsional rigid of the EGMC, the maximum coil efficiency, the maximum magnetic field density inside the coil, the maximum cooling efficiency of the water-cooling cavity, and the maximum magnetic field density in giant magnetostrictive material (GMM) by reducing the magnetic reluctances. The variables needs optimization included the size of GMM, the magnetic permeability of all magnetic materials, the structure of magnetic flux path, coil, and water-cooling cavity. All optimal parameter ranges were determined by application demands. The finest parameters of giant magnetostrictive components were obtained by non-dominated sorting genetic algorithm (NSGA) with space search method, which were proved by the experiments and magnetic field simulations.