<p>This paper introduces a dynamic fitness-distance balance-based metaheuristic algorithm inspired by a physics-based algorithm to optimize the parameters of a hybrid active power filter. The challenge lies in meeting system requirements for voltage and current harmonic distortion levels. The proposed algorithm integrates a dynamic fitness-distance balance learning strategy, historical information, and a nonlinear adaptive weight for effective exploration and exploitation balancing. We evaluate the proposed algorithm on forty-two benchmark problems of two benchmark test suites, comparing results with state-of-the-art algorithms in rankings, convergence, statistical validation, and computational complexity. Furthermore, we apply the proposed algorithm to two hybrid active power filter model case studies, comparing outcomes with existing algorithms. Experimental results show the proposed algorithm’s enhanced accuracy, convergence rate, search capability, and stability across optimization problems. The overall ranking of the algorithms suggests that the proposed algorithm has outstanding potential to deal with complex optimization problems.</p>