Nonalcoholic fatty liver disease (NAFLD) is a prevalent and growing pathology worldwide, presenting a major health concern due to its association with diabetes, obesity, and metabolic syndrome. Early detection and accurate diagnosis of NAFLD are critical for timely intervention and disease management. Ultrasound imaging has emerged as a noninvasive and cost-effective modality for NAFLD detection, and recent advances in machine learning (ML) techniques, particularly convolutional neural networks (CNNs), have shown promise in automating the analysis of ultrasound images. In this chapter, we implement a transfer learning-based approach using 11 different pretrained CNN models for NAFLD diagnosis with ultrasound images. We compare the performance of various classifiers, including k-nearest neighbors, support vector machines, random forests, artificial neural networks, long short-term memory, and bidirectional long short-term memory, for NAFLD detection. Our results demonstrate the effectiveness of CNN-based transfer learning in automating the diagnosis of NAFLD, providing a potential solution for accurate and efficient disease detection and management in clinical practice.