Abstract: In this paper, an improved ant colony algorithm based on the cost of energy consumption and user is proposed to solve the problem of load imbalance in the process of cloud resource allocation. First, the energy consumption cost function model of users and operators is established in the cloud environment. Secondly, we have introduced the dynamic adjustment factor DRF in pheromone update mechanism, which makes the algorithm achieves load balancing state after several iterations. Finally, the simulation test is carried out on the cloud simulation platform named Cloudsim. The experimental results show that the algorithm can reduce the task execution cost, and meet the resource load balancing well.