The main objective of the doctoral thesis is to investigate algorithms for minimizing energy consumption, neural networks, and develop a method for reducing energy consumption in unmanned vehicles. The doctoral thesis consists of 4 chapters, conclusions, and a list of references. In the first section of the doctoral thesis, unmanned vehicles are studied, and the control structure of unmanned vehicles is described. The control of unmanned vehicles with the implementation of the developed method into the existing control structure is proposed. The structure of the newly developed self-learning optimization controller with a neural network is described. The second section of the doctoral thesis describes the spatial models of vehicles, identifies the objective function and energy consumption function. Mathematical models of vehicle motion, neural network, and mechanical model of vehicles are developed. In the third section, a general algorithm for optimizing the energy consumption of unmanned vehicles is developed. Four minimization algorithms are described, two of which are stochastic and two are deterministic. The neural network training algorithm is also described. A self-learning algorithm for optimal energy consumption with a neural network is developed. The fourth section of the doctoral thesis presents the results of the research experiments. The developed algorithms are simulated on computer models. Experimental devices are constructed and used to test the investigated algorithms. Results of the work: The method allows for a 12.93 % reduction in energy consumption by autonomous electric vehicles. The doctoral thesis consists of 126 pages, including an introduction, 4 chapters, 84 figures, 8 tables, conclusions, and 92 references.