Abstract: Mobile edge computing (MEC) designs a capacity-constrained edge server (ES) deployment optimization strategy based on network access point (AP) distribution in wireless metropolitan networks for latency-sensitive computing tasks, overcomes node initialization stability, balances the load of each server, and optimizes the deployment of ES final position using the mean shift algorithm to reduce the communication distance and achieve the purpose of reducing latency. The server clustering deployment is evaluated for different scenarios, and the experiments show that the deployment of server nodes using the residual computational capacity spatial mean clustering (RCCsmc) algorithm has lower latency, more balanced load, and wider applicability scenarios.