The work aims to address issues in product image evaluation caused by the butterfly effect, such as biases in evaluation results and inaccurate selection of case databases, and propose a method for product image styling evaluation based on a data fusion model. Firstly, the research sample and perceptual image were determined, and a survey questionnaire was designed based on the semantic differential method to conduct a survey on the perceptual image of the product and explore users' perceptual needs. Secondly, the simulated annealing algorithm was used to optimize the parameters of the improved support function, and a data fusion model for users' perceptual needs was constructed. Then, the mean value of users' perceptual needs data and the data fusion value were used as inputs for comparison, and the evaluation of product image styling was conducted to guide the selection of the optimal solution. Finally, the proposed method was validated with a case study using a mouse as an example. The product image styling evaluation method based on the data fusion model is more accurate and objective compared with traditional average value data analysis. It effectively captures users' emotional needs, reduces errors in product image evaluation, and guides the selection of a typical case database.