Abstract: Individual feed intake in cattle and sheep serves as a fundamental physiological indicator for assessing animal health status, formulating precision diets, guiding genetic selection, and optimizing feeding management strategies. Traditional monitoring methods such as manual weighing and indicator techniques, while historically valuable, are constrained by significant limitations including high labor intensity, substantial operational costs, poor temporal resolution, and potential disturbance to natural feeding behaviors—often inducing animal stress. Although suitable for controlled research environments, these conventional approaches are increasingly inadequate for meeting the demands of modern livestock production characterized by large-scale operations, precision management, animal welfare, and intelligent automation. The rapid development of animal wearable sensors, machine vision, Internet of Things technologies, edge computing, and deep learning algorithms has catalyzed a paradigm shift in monitoring capabilities. Intelligent perception-based methodologies have emerged as a prominent research focus, enabling revolutionary approaches for real-time, continuous, and automated monitoring of individual feed intake in cattle and sheep. These technological innovations encompass automated feeding stations, acoustic sensors, accelerometers, pressure detection systems, electrophysiological signal monitors, and computer vision systems. The feeding process involves a complex sequence of distinct behaviors including biting, chewing, swallowing, rumination, rumination chewing, and rumination swallowing—all demonstrating significant correlations with actual intake volume. These behavioral manifestations produce measurable changes in feed mass and volume, generate characteristic acoustic signatures during mastication, create distinctive head movement patterns, induce jaw pressure variations, elicit muscle electrophysiological activities, and exhibit identifiable visual feeding actions. Such multi-modal behavioral signatures can be effectively captured by complementary sensing technologies, forming a comprehensive foundation for intelligent intake estimation through advanced data fusion and analysis techniques. This paper provides a systematic and comprehensive review of recent advances in intelligent monitoring technologies for feed intake in cattle and sheep. We present a detailed comparative analysis examining the underlying working principles, reported measurement accuracy, optimal application scenarios, and inherent limitations of each technological approach. Automated feeding stations determine intake through precise measurement of feed mass before and after consumption, offering high measurement accuracy and minimal animal interference while requiring no complex algorithmic modeling, though their high equipment costs and maintenance demand limit application primarily to confined housing systems. Acoustic monitoring technology detects and analyzes feeding sounds generated during chewing and swallowing processes, featuring compact design, easy installation, and minimal animal stress, yet remaining vulnerable to environmental noise interference. Acceleration monitoring technology quantifies intake by analyzing head movement kinematics and integrating jaw motion acceleration patterns, providing cost-effective monitoring solutions with reasonable accuracy. Pressure monitoring technology identifies feeding behaviors through jaw pressure waveform analysis, demonstrating strong anti-interference capabilities but facing challenges in accurately quantifying intake intensity. Electromyographic monitoring technology predicts intake using electromyographic signals from jaw muscles, offering good measurement accuracy but encountering practical limitations in maintaining reliable electrode-skin contact during long-term deployment. Vision-based monitoring technology estimates intake by detecting changes in feed volume or analyzing behavior from image data, delivering high accuracy non-invasively but requiring substantial infrastructure investment. Despite remarkable technological progress, the transition from research prototypes to reliable, widely adopted operational systems faces several critical challenges. This review thoroughly examines persistent issues including excessive energy consumption and limited battery endurance—particularly problematic in extensive grazing systems; the imperative to safeguard animal welfare by minimizing device-induced stress and behavioral disruption; limited algorithmic generalization capability across diverse animal breeds, production stages, and environmental conditions; and difficulties in effectively managing, processing, and extracting actionable insights from multi-source big data. Future research directions should prioritize the development of low-power sensing technologies and edge computing solutions, enhancement of algorithmic robustness for cross-scenario deployment, and construction of integrated data analytics platforms for whole-farm management optimization. This review synthesizes current technological advancements, clarifies performance trade-offs, and identifies critical research priorities, aiming to provide valuable technical references and practical guidance for promoting precision feeding management, enhancing genetic breeding programs, and advancing smart farm management in the production systems of cattle and sheep.