Purpose/Significance To promote the upgrading of the medical insurance audit model in hospitals, so as to relieve management pressure and ensure the safety of medical insurance funds.Method/Process Based on the self-developed medical insurance decision support system and intelligent medical insurance audit system, a new model of human-machine collaborative medical insurance audit is established. Combined with publicity and training as well as data feedback, it guides the improvement of clinical behavior.Result/Conclusion The rate of omissions in medical insurance diagnoses in inpatient medical records has dropped from 28.89% to 3.22%. The scope of audit has been expanded from key medical records to all medical records. The daily manpower input for medical insurance audit has been reduced from 7 full-time employees to 1 part-time employee, and the same-day medical insurance audit rate has increased from 92% to 97.5%. The human-machine collaborative model has enhanced the quality, efficiency and patient experience of medical insurance inpatient medical record audit, providing an effective path to address the pain points in hospital management.