Against the backdrop of the national innovation strategy and the digital transformation of education, the traditional “extensive” training model for innovation and entrepreneurship talents struggles to meet the personalized development needs of students, making an urgent shift toward precision and intelligence necessary. This study constructs a four-dimensional integrated framework centered on data, ”Goal-Data-Intervention-Evaluation”, and proposes a data-driven training model for innovation and entrepreneurship talents in universities. By collecting multi-source data such as learning behaviors, competency assessments, and practical projects, the model conducts in-depth analysis of students’ individual characteristics and development potential, enabling precise decision-making in goal setting, teaching intervention, and practical guidance. Based on data analysis, a supportive system for personalized teaching and practical activities is established. Combined with process-oriented and summative evaluations, a closed-loop feedback mechanism is formed to improve training effectiveness. This model provides a theoretical framework and practical path for the scientific, personalized, and intelligent development of innovation and entrepreneurship education in universities.