The key to achieving TCM syndrome differentiation and treatment through artificial intelligence lies in the standardization，normalization，dataization and intelligence of TCM four diagnostic methods. It analyzes the current status of artificial intelligence in the four examinations of inspection，listening and smelling，inquiry and pulse taking and palpation. Inspection adopts high-definition cameras to capture the entire face，tongue images，sublingual veins，lip color，eyes and other parts. The facial segmentation algorithm，color space model，spectral chromaticity and wavelength reflectance are combined to provide ideas for quantitative analysis of the observation. Inspection collects abnormal sounds such as cough and bowel sounds，and abnormal odors such as mouth，sputum and stool，and uses BP and SVM neural networks and electronic nose and other technologies for qualitative and quantitative analysis，providing a new method for the diagnosis of pulmonary and gastrointestinal diseases. Inquiries use the deep learning（DL）model for human-computer interaction to build a many-to-many probability map of pathogenesis and syndromes，providing probability statistics for accurate prescriptions. Pulse taking and palpation have been upgraded from a primary pressure sensor to a pulse wave touch sensor，which can collect three-dimensional pulse maps and dynamically detect blood flow imaging，diversifying the objective indicators of pulse diagnosis. There are still many problems to be solved urgently in the realization of artificial intelligence for the four diagnoses of traditional Chinese medicine：the standardization and quantification rules for the four diagnoses have not yet been improved；the hardware equipment for the individual collection of the four diagnoses is blooming in a hundred ways，but the hardware equipment for the combined collection of the four diagnoses is in its infancy，and the intelligence of the four diagnoses is low. The degree of integration is low；intelligent models，algorithms and big data are still immature in the field of medical video disease identification；big data information security risks lie，and there is a lack of crosscutting talents between medical and industrial departments to lead the development of the industry. This paper discussed and analyzed these issues.