Abstract: In the context of the development trend of the integration of excellent traditional culture and the intersection of technology and fashion, to address issues such as cultural deficiency and supply-demand mismatch in the development of leather clothing, this paper explored the design process and methods of leather clothing that combine perceptual imagery with artificial intelligence generated content technology. Taking the ancient Chinese Duanyao robe as an example, we extracted users' perceptual imagery of its shape and formulated a leather clothing design plan that integrates user needs, traditional cultural design elements, and technological fashion. First, the psychological cognitive measurement method was used to obtain the perceptual imagery data of the shape of the Duanyao robe, and the factor analysis method was applied to extract the shape imagery factors of the Duanyao robe. Second, the quantification theory type I was used to construct a mapping model of the shape imagery vocabulary of the Duanyao robe and its visual form elements. Finally, guided by the imagery mapping model, prompt words were formed and input into Midjourney to generate leather clothing design plans, and the feasibility of the method was verified. This approach realizes a more cultural, objective, and logical design process and a more targeted design result, providing reference and guidance for improving the design efficiency of leather clothing.