At the present stage of development of dairy farming, the main task facing the industry is the effective management of all processes that ensure the operation of a farm with a large livestock. This task will not be achievable without the use of digital intelligent production control and management systems, tracking the health of animals and their productivity. There are many systems that integrate into production and allow you to monitor the health of cows on a dairy farm. These systems are based on the identification, tracking and collection of information about the physiological state of animals through various sensors. However, all these systems are expensive, demanding to maintain, often fail due to aggressive operating conditions. The article considers the possibility of using an intelligent video monitoring system to analyze the physiological state of cows and predict its changes. To substantiate the requirements for an intelligent video monitoring system of the physiological state of cows, a mathematical model of the life of productive cows in the form of a graph of states is proposed. This method graphically displays the possible states of the system and their possible transitions from state to state. Having constructed a graph of states, relying on the central probability theorem for event flows, it is possible to determine the probability and intensity of state transitions using differential equations. For the functioning of the video monitoring system for analyzing the physiological state of cows and predicting its changes, the proposed mathematical model can be the basis of the system and be based on the collection and analysis of video information about animal behavior. When covering the entire territory of the animal's living space with video cameras, the intelligent video monitoring system will be able to provide early diagnosis of changes in the physiological state of the observed animals and issue a warning, which will allow for timely therapeutic and preventive measures.