Abstract: In the era of the Internet of everything, the huge transmission delays between massive devices and data, clouds and devices have brought multiple challenges to developing applications, processing data, and ultimately improving the quality of service(QoS). As a result, edge computing architectures that significantly reduce the amount of data transferred to cloud servers and the response time of service requests by deploying computing power near end devices have emerged. However, load balancing, security, and mobility of edge devices are still the key points that affect the quality of service of edge computing architectures. In order to solve the above problems, we propose a two-stage QoS optimization scheme to solve the service deployment problem under the mobile edge computing architecture. In the first stage, we consider the decoupling of services and the load balancing of edge servers, model the problem of service deployment, propose a real-time decoupling scheme for centralized services, and design a static deployment strategy; In the second stage, we consider the mobility of edge devices, design a dynamic deployment strategy for device mobility-aware services, and optimize the quality of service under the mobile edge computing architecture. The experimental results on two datasets show that the static deployment strategy proposed in this paper can reduce the response time of service requests by 36%, and the dynamic deployment strategy can further reduce the service response time by 13%.