A method and an apparatus for realizing the load balance of resources in a data center are provided. The method in the technical solution includes: obtaining the current utilization ratios of the attributes of each physical machine in a scheduling domain, and determining the physical machines whose remaining capacity satisfies the current assigned task according to the current utilization ratios, the attributes including the load of the Central Processing Unit (CPU), the load of the memory and the load of the network; determining the average load values of the attributes of the scheduling domain according to the current utilization ratios, and calculating the load variance values of the physical machines according to the average load values and the predicted load values for the attributes of the physical machines; and selecting the physical machine whose load variance value is the least to deploy the current assigned task. The apparatus includes a selection control module, a calculation processing module and a deployment execution module. The adoption of the solution can solve the problem of the load imbalance of physical servers caused by the provision of user requirements differing from the provision of the physical servers.