Abstract: The penetration rate of distributed photovoltaics in distribution networks continues to increase, and the intermittent and fluctuating characteristics reduce the voltage stability and power balance control capabilities of distribution network automation terminals. Traditional proportional integral derivative control strategies are difficult to effectively deal with nonlinear time-varying problems. Therefore, this paper designs a new optimization technology based on improved adaptive self disturbance rejection. This article takes the distribution network automation terminal as the control carrier, the photovoltaic grid connected inverter as the execution object, and dynamically matches the bandwidth of the extended state observer with the error state through the gain adaptive law. At the same time, the improved particle swarm optimization algorithm is used to perform on line rolling optimization on the controller gain, solving the problem of insufficient adaptability of fixed parameters. Furthermore, this article introduces photovoltaic power prediction feedforward compensation to enhance the pre adjustment capability for power transients, forming a progressive control architecture of “parameter self- regulation performance online optimization disturbance feedforward compensation”. The experimental results show that the technology proposed in this article has excellent ability to suppress voltage fluctuations at the grid connection point, laying an important foundation for the safe and stable operation of the distribution network automation terminal system.