With the advent of cloud application, cloud computing has become a charged network service, with task scheduling and resource allocation being the most important issues for a cloud service center. In most algorithms, the requirement for QoS (Quality of Service) has not been considered comprehensively. The objective of this paper is to optimize task scheduling and resource allocation, using an improved ant colony optimization algorithm based on the proposed cost and time models in cloud computing environment. An integrated multi-objective efficiency function is proposed to balance the consumptions for resources allocation. An improved ACO (Ant Colony Optimization) algorithm is proposed for use in cloud environment. By means of entropy, the algorithm's convergence rate could be improved in measuring the uncertainty of cloud resources and updating the global pheromone. Besides, it is demonstrated that better resource balance can be obtained during tasks' execution. Experimental results verify the effectiveness of the proposed improved-ACO algorithm, in convergence, stability, and solution diversity.