Conventional ultrasonography often struggles to accurately differentiate between benign and malignant thyroid follicular tumors, which relying heavily on postoperative pathological diagnosis. Recent advancements in novel ultrasound technologies and artificial intelligence(AI) have shown significant potential in improving diagnostic accuracy, reducing unnecessary surgeries, and decreasing misdiagnosis rates. Emerging ultrasound modalities, such as superb microvascular imaging, contrast-enhanced ultrasound, and ultrasound elastography, provide new approaches for preoperative differentiation of thyroid follicular tumors. This review summarized and discussed the application value of these novel ultrasound techniques and various AI-based modeling methods in the preoperative diagnosis of thyroid follicular tumors, aiming to provide a scientific basis for clinical decision-making.