Virtual machine relocation has attracted attention to enable various cloud computing services and to facilitate virtual machine (VM) migration. Most of existing virtual machine relocating strategies usually focuses on the optimization of single objective in cloud computing environment. But the objective often sacrifices other objectives to achieve the optimal effect. Therefore this paper presents a virtual machine relocating method with multi-factor tradeoff, in which resource utilization is converted into energy and there is a tradeoff between energy consumption and performance at the same time. The strategy effectively avoids the unnecessary migration by using autoregressive model to predict the performance of the future period when choosing the virtual machines to be relocated. The experiment results show that the presented strategy can improve the response time of virtual machine, the energy consumption, and the number of virtual machine migration under the condition of guaranteeing certain SLA (Service-Level-Agreement).