In order to improve the efficiency of the mechanical product surface coating texture design, the design method based on kansei engineering was studied in depth. Firstly, based on kansei engineering, the texture design method and process and key technology of mechanical product surface coating from the theoretical level was proposed. Secondly, for key technology, typical texture samples and typical texture visual imagery vocabulary were collected, conduct visual imagery sensitivity experiments to obtain the relative membership degree of typical texture samples and typical visual imagery vocabulary; Using BP neural network, the “visual imagery-texture parameter” correlation model of mechanical product surface coating was constructed; Finally, take the digital printing machine for example, when the texture parameters were “coating roughness 0.2, coating refractive index 1.18, coating thickness 1.2”, the texture visual imagery values were “safe 3.132, simple 2.657, beautiful 3.089, stable 2.756, solid 2.842”. The results show that the “visual imagery- texture parameter” correlation model had a better subjective trend and can meet the visual imagery needs of users. The method of mechanical products surface coating texture design was maneuverable and scientific, which can provide reference for the design of mechanical product surface coating texture.