An optimized scheduling algorithm in cloud computing environment is designed based on the QoS constraints. Under the QoS requirements with the constraint conditions marked by the SLA(Service-Level Agreement) parameters, tasks are graded to form the priority queue. SLA service parameters are marked by the binary weights to avoid overlap of the Job Level. The method of resource level queue is adopted and the work is reasonably assigned to nodes when allocating computing resources to tasks. Correctness and robustness of the optimized algorithm are verified by experiments results comparison with several algorithms.