<p>Optimal coordination of directional overcurrent relays is essential for modern power systems to take fast actions against faults without causing vulnerabilities in the selectivity of protection systems. However, due to the nonlinear and highly constrained structure of this problem, it may not always be possible to achieve a desirable level of solutions. Furthermore, the problem becomes more challenging and time-consuming as the complexity of the protected system increases. At this point, to address the aforementioned issues, there is a need for effective optimization methods in terms of solution quality and computational efficiency. In this way, a highly promising method can be achieved for protection coordination applications in both offline and online frameworks, especially by leveraging supercomputing technologies. In this manner, this paper proposes a new variant of walrus optimizer called improved walrus optimizer (IWO) to solve this challenging optimization problem. Even though the WO has a strong exploitation ability, it can fail in avoiding local optima due to its weakness in exploration ability. To address this shortcoming, chaotic initialization and oppositional mutation procedures are introduced to the original WO algorithm. In addition, a modification to the position updating rule of the exploration phase is made to enhance the algorithm’s convergence performance. The results performed on the 8-bus, 15-bus, IEEE 30-bus, and New England 39-bus test systems prove the superiority of the IWO algorithm.</p>