Abstract <p>Managing the dynamic nature of flying ad hoc networks (FANETs), comprising a group of Unmanned Aerial Vehicles (UAVs) requires an adaptive clustering strategy that ensures network stability and efficient communication among the UAVs. This paper introduces a novel framework, EnRSAF, based on a chaos-driven Reptile Search Algorithm (RSA) with weighted optimization and hybrid mutation, for efficient FANET clustering. To address the inherent RSA challenges of local optima and premature convergence, an enhanced version (EnRSA) is devised and adapted for the FANET clustering. EnRSAF enhances cluster stability and ensures reliable data delivery by forming stable UAV clusters that minimize frequent reconfigurations and maintain consistent communication. It incorporates a multi-criteria approach to identify the best CHs, focusing on improving cluster lifetime and energy efficiency. Additionally, an effective route selection function is designed to adapt dynamically to changing network conditions, thereby reducing packet losses and ensuring timely data transmission. With these novel enhancements, EnRSAF establishes itself as a reliable solution for FANET applications. Comprehensive simulation results reveal significant improvements across metrics like cluster lifetime, energy efficiency, and packet delivery performance, when compared to recent techniques.</p> Graphical Abstract <p></p>