Ruminant livestock is an important source of meat, milk and other food for human beings. With the improvement of people’s requirements for the output and quality of ruminant livestock products, the traditional manual supervision mode, which is time-consuming, labor-intensive and high labor cost, has been difficult to meet the needs of large-scale ruminant livestock breeding. Ruminant livestock behavior contains a lot of body condition information. The intelligent monitoring of ruminant livestock behavior is helpful to identify abnormal behavior of ruminant livestock earlier, evaluate the health level of ruminant livestock, early warning of abnormal physiological state of ruminant livestock, and assist farmers to adjust breeding strategies in a timely manner to achieve low cost-effective, efficient and profitable production process. Firstly, the monitoring methods for basic movements (lying, walking and standing), rumination, eating and drinking, lameness of ruminant livestock were overally described. Secondly, the different characteristic indicators to identify the condition of ruminant livestock in estrus, parturition, disease and pain were analyzed in detail and the physiological condition identification method was introduced based on the characteristic indicators. Thirdly, the problems and challenges of ruminant livestock behavior monitoring methods were summarized. Finally, the future development directions of relevant key technologies were prospected, including optimizing sensor power consumption, fusion of multi-sensor data, reducing data transmission delay, reducing large-scale data annotation, lightweight deep learning models and deep analysis and application data.