Plasma electrolytic oxidation, an advanced surface treatment method, is capable of generating a robust ceramic coating on the exterior of materials utilized in electrical engineering. However, the complexity of the plasma electrolytic oxidation process, coupled with the continuous variation of load parameters as the reaction time increases, poses challenges to theoretical research, and a comprehensive explanation has not yet been provided. In response to this issue, an online identification method for load parameters in plasma electrolytic oxidation based on the recursive least squares algorithm is proposed in this paper. An electrical equivalent circuit model of the plasma electrolytic process was constructed, and based on this model, the load resistance and capacitance parameters were estimated online using the recursive least squares algorithm. The experimental results show that the proposed method can accurately track the trend of changes in load parameters during the plasma electrolytic oxidation process, providing an effective technical means for real-time monitoring and quality control of the surface treatment process of light metal materials.