The object of research is the decision-making process in an intelligent video surveillance system based on neural networks. The subject of research is models, methods, technologies of data recognition in an intelligent video surveillance system based on neural networks. The goal of the work is to investigate the features and prospects of using video surveillance systems using neural networks. Research methods – image segmentation methods, neural network training methods, optimization methods. The qualification work is devoted to the analysis of approaches to detecting anomalous human behavior in real time in a video sequence and the analysis of modern applied pattern recognition systems. The work explores the design of the architecture of an intelligent video surveillance system, and neural networks for image processing in intelligent surveillance systems.