The system of automatic control of the productivity of the PGV-1000 steam generator of the power unit of the nuclear power plant with the WWER-1000 reactor is considered. This control system includes a system for automatically controlling the water level in the steam generator. The error signal, which is fed to the input of the productivity controller, is generated by the signal of the specified speed of the drive turbine, as well as pulses from the thermal power sensor and from the water and vapor pressure sensors. Based on the model of the water level control system in the steam generator, mathematical models of performance control systems have been developed under various control laws. To formulate optimization tasks for performance management systems, vectors from variable parameters of performance controllers are formed, constraints on these parameters are imposed, constraints are created for the areas of system stability, direct indexes of system quality are introduced, and a vector objective function is created that takes into account all imposed restrictions and quality criteria. Computational experiments on optimizing the controllers of productivity control systems were performed on the basis of direct quality indexes of systems by modified methods of step adaptation and Nelder – Mead. Analysis of the results of optimization of performance control systems allows us to conclude that, in comparison with the minimization of integral quadratic estimates, the optimization of direct quality indexes has made it possible to substantially improve the speed of the systems. The greatest value of the control time takes place for the differential controller, and the smallest identical values – for the proportional-integral and proportional-integral-differential regulators. Theoretically, the expediency of using a more simple proportional-integral controller in the steam generator productivity control system is justified. Additional analysis of the state variables in the productivity control system and comparison of transients before and after optimization allows us to conclude that the maximum deviation of the water level in the steam generator has decreased and its oscillations have disappeared. Also, fluctuations in water flow disappeared, the maximum deviation and oscillation of the control error decreas ed substantially, the overshoot and oscillations of other state variables practically disappeared. Thus, the optimization of the quality indexes of the automatic control system of the steam generator’s productivity has made it possible to significantly improve the main processes in it.