This study aims to solve the problems of high threshold and difficult to unify the style of traditional budding character design with the help of generative artificial intelligence (generative AI), to meet the demand for personalized creation, and to promote cross-disciplinary integration and cultural dissemination. The study adopts AI models based on Generative Adversarial Network (GAN) and Variable Autocoder (VAE), uses stable diffusion as a tool to generate moe character images; classifies and encodes moe elements according to manifestation, environment, and spiritual elements, and decodes them through quantitative research and user analysis; and conducts questionnaire surveys to obtain the needs of the group of people aged between 15 and 40. The research results are verified by two groups of design practice, one is generative artificial intelligence assisting manual optimization to meet the demand of personalized design, and the other is generative artificial intelligence assisting cross-cultural fusion design, both of which have achieved good results. The study shows that generative AI can efficiently generate various budding characters, optimize the details by combining human-computer collaboration, and promote design standardization and innovation. With the development of technology, generative AI will play a greater role in the field of creative design, helping design to personalization and diversification, and injecting new momentum into the development of cultural industry.