Abstract: To meet the high precision and efficiency requirements in medical consultation scenarios, we adopt the key technologies of human-machine computing to achieve a deep integration of manual and machine consultations. This paper explores human-machine computing techniques at different stages of medical consultations, including human-machine task allocation, human feedback enhancement, and human-machine decision-making complementarity. First, we optimize task allocation using deep reinforcement learning and employ a hierarchical learning framework to achieve efficient human-AI collaboration. Then, we introduce a consultation enhancement mechanism based on human feedback to leverage expert knowledge and improve AI performance. Furthermore, considering the complementary characteristics of human and machine intelligence, we propose a medical diagnosis framework based on human-machine combined advice to enhance the diagnostic performance of human-machine medical teams. Finally, we discuss the development trends and challenges of human-machine computing in medical consultations and highlight its potential value in intelligent healthcare.