The issue of scheduling in a cloud is an NP-hard optimization problem. Maintaining a load balance between processing units in the system is of great significance in cloud technology. When a set of tasks arrive at the cloud, the system is supposed to respond to all of them as it manages to achieve the shortest possible time. This paper uses Cuckoo algorithm to perform such a management. The purpose of the proposed method is achieving an order of processing units such that the time of responses to queries is minimized. The input to cuckoo algorithm is the number of virtual machines and the number of tasks. By examining various orders of these machines, the proposed method allocates hosts to tasks in a proper way. Simulation results show that using Cuckoo algorithm for the intention of reaching the best order of processing units can lead to improved performance parameters. In addition, simulations reveal that if tasks are scheduled without any primary information about the resources, the results will not be satisfying enough.