The optimal placement of edge servers is fundamental to enhancing service delivery in edge computing, yet it presents a complex trade-off among conflicting criteria such as communication latency, load distribution, and energy efficiency. To address this multi-faceted challenge, this paper introduces a hierarchical optimization framework. Initially, the framework leverages an enhanced Density Peak Clustering Algorithm (DPCA) to intelligently partition base station networks into viable deployment regions. Following this, a novel Multi-Objective Particle Swarm Optimization (MOPSO) algorithm determines the optimal quantity and precise locations of edge servers within these designated regions. Our MOPSO variant is distinguished by an adaptive velocity update mechanism guided by Pareto dominance and a local search operator to improve solution diversity. We validated our framework’s efficacy through extensive simulations on a real-world dataset featuring telecommunications base station topology and traffic data. Comparative analysis against established benchmarks, including Top-K, K-means+PSO, and genetic algorithms, demonstrates the superiority of our approach. It achieves significant reductions in communication latency by 15.8%, improves load balancing by 18.4%, and lowers energy consumption by 12.3%, while also producing a more comprehensive and well-distributed Pareto frontier. These findings underscore the strategy’s effectiveness for practical, large-scale edge infrastructure deployment.