Relevance. The need to improve the energy efficiency of oil production wells operation by rod deep installations. Taking into account the trend of annual growth in the average water content of oil production wells, an increase in the share of hard-to-recover reserves and a complicated fund, the practical implementation of methods to reduce operating costs is the most important task of the modern Russian oil industry. In this aspect, intellectualization of control systems, information processing and monitoring systems allow increasing the energy efficiency of individual elements of the unit, without changing the technological parameters of the unit as a whole. The electric drive and its control systems and kinematic values of the pumpjack can be considered as separate elements of optimization. It is difficult to imagine applying of new solutions in the kinematics of the pump unit that differ from the generally accepted ones. Considering the drive control system, there are limited opportunities to improve energy efficiency, because of limited speed control range. On the other hand, the parameters of the rotor counterweight significantly affect the active electric power consumed by the electric drive. Thus, the development of algorithmic solutions for intelligent control stations of pump units becomes an issue of the day. Aim. To form the author's methodology for balancing a rocking machine with a rotary counterweight and checking its operability in laboratory conditions. Methods. Research methods are based on the use of the provisions of modern control theory, automated electric drive theory, optimization theory, and simulation methods. Results. The authors have proposed the method of finding the optimal balancing of the rocking machine. The authors carried out the theoretical calculation of the optimal balancing for the target oil well of the Unvinskoe oil field and a full-scale simulation of this process on a test bench specially developed on the basis of Perm National Research Polytechnic University. As a result of checking the operability of the developed technique, a decrease in the root mean square torque value by 10.32% per swing cycle and a decrease in active energy consumption by 7.15% were recorded on the test bench. Conclusion. The practical application of the developed algorithm will allow achieving a significant reduction in the power consumption of the rocking machine. The developed algorithm can be implemented as a function of a frequency converter of a drive electric motor of a rod borehole pumping unit.