<p>Cloud computing systems work in environments that are always changing, so they need to manage resources well to support a variety of virtual machines and workloads that change. It is an NP-hard optimization problem to move and place virtual machines across distributed cloud infrastructures while keeping costs low and performance high. Nature-inspired metaheuristic algorithms can effectively navigate large and complex search spaces. However, many current scheduling methods rely on fixed migration thresholds and a singular optimization strategy, which restricts their adaptability to changing workload conditions. This paper presents a Hybrid Jellyfish–Dragonfly Optimization (HJDO) methodology integrated with a Fuzzy Inference System (FIS) for the adaptive management of virtual machine migration. The FIS changes migration thresholds on the fly based on how the workload behaves in real time, getting rid of the problems with static fuzzy decision rules. The main thing this study adds is the closed-loop interaction between hybrid optimization and adaptive fuzzy control, which makes it possible to keep improving scheduling decisions. The optimization process includes a multi-objective fitness function that takes into account energy efficiency, resource use, and cost effectiveness. The suggested Enhanced Dynamic Fuzzy Load Balancing (EDFLB) scheme has been put into action and tested against the OFLB and FCFS scheduling methods. Experimental assessment over several iterations, corroborated by statistical validation employing mean values, 95% confidence intervals, and effect size analysis, verifies that EDFLB attains substantial performance enhancements, including a 27.78% increase in makespan and a 44.24% enhancement in average response time, illustrating enhanced scalability and energy-efficient cloud resource management.</p>