The development of digital-intelligent technologies poses challenges to architectural education in terms of core skill transformation and the cultivation of interdisciplinary innovative talents. To address these challenges, this study constructed a "teacher-student-machine" tripartite collaborative teaching model and proposed a five-phase teaching method integrating the architectural design process: task analysis, concept generation, scheme development, digital construction, and critical reflection. A ten-week quasi-experimental study was conducted to compare the design outcomes and competency development between a human-machine collaboration group and a traditional teaching group. The results indicated that the human-machine collaboration model significantly outperformed the traditional model in terms of the innovativeness and expressiveness of design outcomes, design efficiency, critical thinking, and decision-making ability. Intelligent machines served as cognitive enhancers and thought catalysts in teaching, rather than mere substitutes. This study provides theoretical and practical references for the deep integration of digital-intelligent technologies and architectural education, exploring a feasible path for cultivating future design talents.