Network management becomes more difficult for the increase of data transmission speed and network complexity, so the paper presented an intelligent multi-Agent model for virtual resource, described the process of the multi-Agent to virtual resource, and discussed the processing mechanism of different Agent. The proposed model was able to analyze social media resources in real-time by user context and system state. It automatically allocated resources suitable for users according to the virtual resource usage type and the information demand analysis of user context. The model was evaluated by dynamic scheduling method of virtual resources in cloud computing and the MovieLens system. The results show that the proposed model has better performance, can achieve the dynamic scheduling and load balancing of virtual resource, so that users can utilize efficiently virtual resource in the cloud computing.