The article analyzes the problems and methods of research, development, and implementation of control theory and machine learning in industrial corporation. A model of organization of corporate production is built and studied, at the top level of which is the Principal, at the middle - the Representative and at the bottom — the Agent. The interaction between the Principal and the Representative was studied, and the features of their strategic behavior were identified. An optimal mechanism of control of the Representative on the part of the Principal, aimed at revealing the production potential, is proposed. The nature of the interaction between the Representative and the Agent and their strategic behavior is determined. The procedures of reinforcement learning of the Representative and motivation of the Agent are considered. An optimal mechanism of control of the Agent on the part of the Representative is developed. Based on the obtained results, a comprehensive mechanism for corporate management has been formed, which motivates active elements of an industrial corporation to increase production. The obtained results are applied in the service management system of Russian Railways corporation.