This article explores models and methods of information transmission essential for the coordination and management of UAV swarms in dynamic environments. Various communication architectures, including centralized, decentralized, and multi-layered approaches, are analyzed in terms of reliability, scalability, and fault tolerance. A classification of routing protocols, such as topology-based, geographic, and swarm intelligence-based, is presented, along with a discussion of their advantages and limitations. Special attention is given to optimization methods, including predictive and multi-path routing, which demonstrate high efficiency in networks with high mobility and dynamic topology changes. The study also includes experimental validation through simulations in CoppeliaSim, evaluating different control models (centralized, decentralized, and distributed) based on communication delay, response time, and resilience to node failures. Results indicate that hybrid approaches combining decentralized communication with adaptive routing strategies significantly improve network reliability and energy efficiency. The findings emphasize the importance of integrating machine learning-based routing mechanisms and hybrid control solutions to enhance UAV swarm performance in complex and unpredictable environments. Future research should focus on further refining AI-driven routing techniques to support real-time decision-making and scalability in large swarm networks.