There are no doubts that wireless networks have conquer people everyday life. Electronics manufacturers and scientist each year are increasing transmission data rate. Nevertheless, existing technologies allow using wireless networks in everyday activities without much problems. It is possible to say that human related communication problems are solved. Nowadays more and more projects are implementing machine-to-machine communication. It is treated as the future of networks – the Internet of Things. The demands of these applications are completely different from human communication. As the main requirement becomes low energy consumption and low transmission delays. The PhD thesis focuses on wireless data transmission in time critical applications. Such a systems combine low delay and energy consumption requirements. Wireless sensor networks in industrial applications are used as an example of such system. It makes controllable system to interact with an environment. That makes a formalization of such a system a nontrivial task. Since real environment objects nature is nonlinear and dynamic. Making a wireless network with predetermined time parameters is hard task, which is being solved in this thesis. The thesis proposes the wireless transmission system implementation by using analytical model of a system. By using dynamic parameters, it is possible to predict critical points of a transmission network. That allows lowering system’s design time, since problems are solved on the early stage. To realize system formalization, Hamerstein-like model is used. It allows separating nonlinear part of a system, from dynamic part. To implement data transmission, hybrid network media access algorithm is proposed. It exceeds existing algorithms in the mean of higher transmission frequency and timing stability. Proposed formalization method and media access algorithm were tested in international science projects and real industrial application. The thesis consists of introduction, 5 chapters, conclusions and appendix. It contains 130 pages, 65 figures, and 130 references.