<p>Wireless sensor networks (WSNs) are challenged by three main factors: energy usage, delay, and network lifespan. This paper proposes an enhanced marine predator algorithm (EMPA) to effectively determine cluster heads (CHs) in WSNs, aiming to conserve energy, minimize delay, and extend network longevity. EMPA features a Taylor-based neighborhood mechanism (TNM) to improve local exploitation and a dynamic opposition adaptation (DOA) mechanism to avoid stagnation in local optima. These advancements allow much better exploration and exploitation of the search area, optimizing CH selection according to multi-objective parameters, like energy efficacy, distance minimization, and quality of service (QoS). The proposed strategy incorporates exploration and exploitation dynamically to strengthen solutions in terms of robustness and adaptation to different WSN configurations. The experimental evaluation of EMPA proved that this algorithm outperforms other algorithms by achieving incredible convergence speed, longer network lifetime, and communication efficiency. These results confirm that this algorithm is suitable for practical applications requiring energy-aware optimization in large-scale WSNs.</p>