Against the backdrop of the deep integration of information technology and education, intelligent research and training, as an important avenue for teachers’ professional development, has garnered significant attention for its effectiveness enhancement. This paper delves into the challenges faced during the implementation of intelligent research and training and proposes optimization strategies from an evidence-based perspective grounded in artificial intelligence. The study begins by outlining the core connotations and development trends of intelligent research and training, emphasizing the data-driven and personalized characteristics empowered by AI technology. Subsequently, it systematically analyzes specific issues and difficulties encountered in the current implementation of intelligent research and training, including the depth of technological application, the design of research and training activities, teacher participation and adaptability, evaluation of research and training effectiveness, and the data-driven closed loop. In response to these dilemmas, the study proposes highly actionable enhancement strategies, including constructing an AI-empowered design framework for evidence-based research and training activities, developing intelligent personalized learning support systems, establishing a big data-based learning analysis and precise intervention mechanism, designing a multidimensional AI-empowered evaluation system for research and training effectiveness, and creating a data-driven closed loop for the continuous optimization of intelligent research and training.