With the development of edge intelligence technology, the request of users requires lower response latency. The cloud-edge collaborative computing model provides an opportunity to optimize application response time. However, the edge servers with limited resources cannot achieve the best response quality of all services. Therefore, a reasonable intelligence service placement strategy in electric area is urgently required to maximize resource utilization and improve the quality of services (QoS). Furthermore, the difference in intelligence service attributes in the electric area and the load dynamics make the static placement strategy insufficient. Therefore, to improve the QoS defined by the average response time of services and the cost of the migration of the service, we propose a service replacement strategy. In this algorithm, we first select the service that contributes to improving QoS from the services that exceed the maximum tolerance time for replacement and then balance the request scheduling delay and the service migration cost to obtain an approximately optimal replacement strategy. Finally, we use simulation experiments to verify the effectiveness of the scheme.