-In the paper the main problems of power control strategies in series hybrid electric vehicles are discussed. Classical regulators are compared with neural networks, their advantages and disadvantages are shown. Experimental data sources for training neural networks and proofing results are suggested and limitations of their application scope have been considered. It is shown that improving the quality of the regulation of electric drives by using a neural network controller can lead to an increase in the reliability of regulatory systems, which subsequently leads to the saving of electricity and resources. Using the experimental data it makes possible to construction and synthesis of a neural regulator. Thus the following research directions are evaluation of optimal neural network structure and modeling its performance.