The present disclosure relates to a method and a device for implementing load balance of data center resources, and the method includes: obtaining current utilization ratios of properties of each physical machine in a scheduling domain, and determining, according to the current utilization ratios, physical machines that have residual capability satisfying a current allocation task, where the properties include a CPU load, a memory load and a network load; predicting a load degree of each of the physical machines on the assumption that the virtual machine is allocated by each of the physical machines; selecting, according to the predicted load degree of each of the physical machines, a physical machine from the physical machines to allocate the virtual machine. By adopting the solutions of the present disclosure, the problem of the load imbalance caused by the inconsistency of specification requirements of users and specifications of physical servers is solved.