This paper summarizes the research progress of artificial intelligence technology in ultrasonic diagnosis of thyroid nodule. In recent years, the incidence of thyroid cancer is on the rise, and thyroid screening is included in routine physical examination, which increases the clinical workload. Oversensitive and oversensitive. Although pathological diagnosis is the gold standard, not all lesions need to be removed at the same time. FNA can effectively improve the accuracy of diagnosis, develop individual diagnosis and treatment, and avoid over-diagnosis. Ultrasound, as the first choice in the diagnosis of thyroid lesions, can effectively find the lesions and preliminarily judge the biological behavior of the lesions. It is convenient and effective, and has unique diagnostic advantages, but it has a certain operator dependence. The development of artificial intelligence provides a "second opinion" for clinical work and becomes a "third eye" for doctors, assisting doctors to find lesions, improving the accuracy of diagnosis and reducing the diagnosis deviation caused by human factors. Although the diagnostic ability of CAD system based on DL is comparable to that of experienced physicians to a certain extent, it lacks transparency and interpretability, and still has a long way to go in information security.