The management of resources and scheduling in cloud computing environment is a complex undertaking mainly due to resource's geographic distribution, heterogeneity and dynamic. A Qos guided task scheduling model, being composed of some scheduling strategies and a QoS guided scheduling Sufferage-min heuristic algorithm, is presented. Taking account for not only the resources' properties of dynamic and heterogeneity, but also the multi Qos levels of both the resources and tasks, this model try to improve the scheduling efficiency by dividing the tasks and resources into two groups of high Qos level and low Qos level and using different scheduling approach respectively. The following experiment shows that the make span value, the key performance, is successfully being shorten.