Biologically inspired design inspires inventions and new technologies in the field of engineering by mimicking functions, working principles, and structures in the biological domain. Aiming at the problem of cross domain knowledge acquisition in the existing biologically inspired design process, a functional semantic clustering based on functional feature semantic correlation and a environment constraint clustering composition based on environmental characteristic constraining adaptability are proposed. Knowledge cell clustering algorithm, and the corresponding prototype system is developed. Finally, the effectiveness of the method is verified by the visual prosthetic device design.