Abstract: The optimal active disturbance rejection control (ADRC) algorithm for paralleling active power filters (APF) is proposed. The improved immune particle swarm algorithm is employed to achieve the global optimization of the active disturbance rejection controller's parameters. The dynamic optimization performance of the immune clonal selection algorithm is further enhanced by the wavelet variation. Based on the immune particle swarm algorithm, a kind of active disturbance rejection control system for parallel APFs is constructed. Experimental results indicate that the proposed control algorithm does not depend on the precise model of controlled plant, and can restrain the negative effect of high frequency interference on the performance of APF compensators. In addition, the effective control for APF is achievable by the proposed algorithm with proved good dynamic characteristic and robustness.