As the Internet continues to evolve, cloud computing has become a hot research topic during recent years. The capability of virtual machine (VM) migration enables a very high utilization of resources in data center. The aim of most live migration strategies is to reduce system overload and improve overall performance. However, these live migration strategies can lead to starvation of VMs with high migration priorities. This thesis proposes some balancing strategies for VM live migration which consider performance for VM migration and handle starvation of VMs. The results of simulation show that our proposed strategies achieve much better performance and have more capability to avoid possible starvation than existing ones.