The paper aims to improve the efficiency of product perceptual design development and the accuracy of image matching, and propose a method of product image design elements combination derivation based on analytic hierarchy process (AHP) and BP neural network by combining qualitative and quantitative methods. Firstly, the database of product image and modeling is established by web crawler and affinity graph method to obtain the image and design elements; Secondly, AHP is used to construct the product hierarchy model and judgment matrix, and the weight coefficients of image and design elements are calculated and analyzed; Then, based on the morphological decomposition method and weight results, the type of design elements and the optimal combination code are obtained. The semantic difference method (SD) is used to obtain the average value of the user perceptual image of the combination code; Finally, Kansei Engineering and AHP-BP neural network are used to build KAB key design elements combination prediction model. Based on this model, the design scheme of quadrotor UAV is predicted, and the technique for order preference by similarity to an ideal solution (TOPSIS) is used to verify and evaluate it. The results show that the combination coding of design elements highly matched with the perceptual image of the target can be obtained through model calculation. Based on this model, the objective and accurate product image modeling design elements combination can be quickly obtained, and the efficiency of product design and development process can be improved.