Significance Through the establishment of a quality control system for electronic medical record (EMR) content,the standardization and normalization of medical record writing is realized, and the quality of hospital medical record is improved. Method/Process The intelligent medical data center is built based on hospital medical data, and the knowledge base and rule base with tumor specialty characteristics are formed by combining natural language processing (NLP) and machine learning technology. The new quality control mode of “pre-audit, comprehensive coverage, process supervision and closed-loop management” of EMR is realized.Result/Conclusion After the application of the medical record quality control system based on NLP, the quality control coverage rate increased from 1% to 100%, and the rate of class A medical records increased to more than 96%, with good real-time and accuracy, providing a solid information foundation for the high-quality development of hospital medical records.