With the wide application and development of imaging technology, the number of accidentally detected thyroid nodules in clinical practice is increasing, and the incidence of thyroid cancer is also on the rise, but the preoperative diagnosis of thyroid nodules has been tricky, and there is no simple and specific method to identify benign and malignant thyroid nodules. Conventional ultrasound imaging has become the preferred method for the diagnosis and differential diagnosis of thyroid nodules because of its ease of operation, low price, high detection rate of thyroid nodules, and safety without radiation. Ultrasound signs such as echogenicity, border, blood flow, presence of acoustic corona and calcification of nodules are usually considered to be important in distinguishing benign and malignant nodules, and are also an important clinical basis for deciding whether a patient needs surgical treatment. However, recently, we have analyzed the preoperative high-resolution ultrasound (≥7.5 MHz Scanner) signs and postoperative pathology of more than 1070 patients with thyroid nodules in our hospital since January 2009, and concluded that there are limitations in the differentiation of benign and malignant nodules by ultrasound imaging. The above ultrasound signs are not decisive in distinguishing benign from malignant nodules: 22% benign and 66% malignant nodules are hypoechoic; 77% benign and 85% malignant nodules have no acoustic halo around them; 16% benign and 29% malignant nodules have unclear borders; 25% benign and 40% malignant nodules have irregular morphology; 21% benign and 57% malignant nodules are calcified, including 13% benign and 47% malignant microcalcifications; 42% benign and 57% malignant nodules have abundant internal blood flow; 4 benign and 4 malignant nodules are associated with swollen lymph nodes in the neck. The number of lymph node swelling in the neck was 4% benign and 21% malignant. Univariate statistics showed that the sonographic features that were significant for the diagnosis of benign and malignant thyroid nodules included cervical lymph nodes, internal echogenicity, microcalcifications, borders, peripheral halo, morphology, and blood flow distribution. Multifactorial logistic regression further suggested that cervical lymph nodes, nodal echogenicity, nodal borders, microcalcifications, acoustic corona, and morphology were the most characteristic. However, none of the signs had both high sensitivity and specificity, and certain signs previously thought to be of greater value, such as nodal microcalcifications, were found in less than 50% of those with malignancy and in up to 13% of those with benignity. Because of the high percentage of benign thyroid nodules, even in this study group benign nodules accounted for 75% (malignant nodules accounted for 25%) and there may actually be little difference between benign and malignant nodules in those with microcalcifications (104 vs. 126 cases in this group). However, ultrasound technology has been rapidly developing in recent years, and the use of ultrasound elastography, ultrasonography, and three-dimensional flow-energy imaging (3D-CPA) provides a new basis for the differential diagnosis of benign and malignant nodules and is a powerful complement to conventional ultrasound imaging. However, these new technologies are not perfect: ultrasound elastography is not very helpful in the diagnosis of thyroid follicular carcinoma; ultrasonographic enhancement patterns of thyroid cancer are diverse, making the differential diagnosis more difficult; 3D-CPA has not yet been able to achieve real-time 3D imaging, and the overall performance needs to be further improved. In conclusion, ultrasound imaging has an important value in the diagnosis of thyroid nodules, but still has great limitations, and ultrasound imaging cannot completely replace thyroid puncture biopsy and other tests.