The traditional education model focuses on students' test-taking abilities, resulting in students lacking the innovation, internal drive, and qualities required of ‘new talents’, and not meeting the standards of “new talents”. In view of the problem, this paper proposes a method based on students' multiple intelligences and character strengths. The multiple intelligence and character strengths mining algorithm based on the fusion model uses the bagging integrated learning method to fuse the three distributed large language models of ChatGLM-6B, ChatGLM2-6B, and Qwen-7B, and proposes mining students' multiple intelligences and character strengths.