In this paper, the concurrent optimization scheme of the exercise health prediction platform based on C/S architecture is reported. First, a cluster of servers is deployed to handle concurrent user requests. Second, based on Nginx, the front-end server was built as the user access port. And third the Spring Boot framework is adopted to deploy the human motion simulation model on the back-end application server. At same time, some optimization schemes about server are designed and implemented to enhance the concurrent processing capability of the cluster. Experimental results show that the concurrent optimization scheme presented in this paper has better performance. The processing efficiency of concurrent simulation prediction request is improved.