When developing new products, brand designers must analyse related products, which is a complicated and time-consuming process. Modern product design often requires complex engineering processes; product development requires extensive knowledge but there is also a demand for shorter product design cycles. Therefore, we propose a method based on extension theory and the analytic hierarchy process for identifying product-related knowledge, to aid the development of new products. First, based on our understanding of extenics, matter-element and relational meta-models of product form, function, and structure are established. Then, we define different primitives of brand identity. Finally, using an "extensional analytic hierarchy process" (EAHP), a hierarchy is established and the weights of different primitives are calculated. Various combinations of primitives are used to facilitate knowledge transfer for computer-aided intelligent design. Design data for multiple cases are analysed to verify the feasibility and effectiveness of the method. The method was verified in a physiological signal experiment, and the results show that the method can effectively accumulate product knowledge. Rapid data mining is important for market competitiveness.