Abstract: Based on the relevant theories and methods of Kansei engineering, the existing robot products were analyzed by semantic difference method, and the user perceptual image evaluation value was obtained by combining the similarity questionnaire. The product perceptual image evaluation system was constructed by using analytic hierarchy process, KJ method, expert demonstration method and semantic difference method.Based on the morphological analysis and modeling feature extraction of the target product, the perceptual image evaluation value is transformed into modeling design elements and applied to the modeling design of medical assistant robot, so as to establish the association between user perception and model elements of medical assistant robot, effectively extracting users’ emotional needs for the product model of medical assistant robot and improving the feasibility and applicability of the product model design of medical assistant robot.