In recent years, themultimedia streaming technology becomes increasingly mature, and as network bandwidth and computing power of personal hand-held devices are developed; therefore, it leads the requirements for multimedia quality increased. In order to provide the quality multimedia content to numerous users, how to divide the loading of content servers to improve the streaming quality is an important challenge. As the concept of multimedia cloud network is formed, how to allocate multimedia streaming service to the nodes in the media cloud network is discussed in this paper. This study designs H.264/SVC streaming service at the media cloud computing center, with the video most suitable for client-side quality provided based on H.264/SVC features (temporal scalability, spatial scalability, and quality scalability) and network bandwidth. In addition, the loading balance and communication mechanisms between nodes are discussed, where the best node is selected by the evaluating node, client-side bandwidth and computing power, for determining the appropriate video streaming path that is used to provide quality multimedia streaming service. According to the experimental results, the bandwidth prediction error rate for general multimedia network streaming service can be maintained at about 6 %, while the utilization rate of various nodes of the multimedia cloud computing center is maintained in a balanced state during the period of executing multi-streaming service.