The increasing requirements on cloud computing entail building up large-scale data centers. How to operate data center in an efficient way is an urgent problem to be solved. Traditional virtual resource allocation strategies didn't take into full consideration how to decrease energy consumption of data center and the time complexity of strategy formation. This paper proposes an energy efficiency virtual resource allocation strategy( EEVRAS) by modeling virtual resource allocation problem as a problem of path construction, improving elitist strategy for ant system(EAS) to optimize resource allocation scheme. Strategy has a lower time complexity. Simulation results illustrate that compared with traditional virtual resource allocation strategy, with the limit of server performance index, EEVRAS can use fewer severs to construct virtual cluster, thus efficiently decreasing the energy consumption of data center.