The article examines the issue of ensuring the security of the Internet of Things (IoT) at industrial facilities, where IoT devices are actively implemented to control, monitor and automate production processes. This is accompanied by an increase in security risks that can negatively affect the functioning of enterprises. Industrial IoT systems that are integrated with corporate networks and the Internet are open to external threats, which could allow attackers to gain access to critical data or even control devices. Among the main security problems, we can highlight the insufficient frequency of software updates for devices at remote sites, a low level of physical security, and the lack of uniform security standards. The main goal of the article is to develop a methodology for comprehensive security testing of industrial IoT systems, which includes the assessment of cyber threats, the development of testing approaches, and the selection of tools to ensure reliable protection. The research focuses on analyzing possible threats to IoT systems, including vulnerabilities in devices, communication protocols, and network infrastructure. As a result, recommendations are provided for strengthening the security of industrial IoT networks using a comprehensive approach to testing, modeling possible attacks, and evaluating physical defenses. The importance of using Wireshark and tcpdump tools to thoroughly analyze network traffic and detect anomalies is also emphasized. The article highlights the main criteria for choosing communication protocols in IoT systems: reliability of transmission, minimization of delay, high level of security, scalability, energy efficiency, error tolerance and compatibility with various technologies. Industrial IoT applications require the use of protocols capable of providing stable data transmission and an appropriate level of protection. Particularly important is also the possibility of modeling the behavior of devices in the network, which allows you to assess risks, check reliability and find optimal scenarios of system operation. For this purpose, the use of Markov models is proposed, which allow predicting the states of devices or network nodes and estimating the probability of their vulnerability to attacks.