There are some key nodes in P2P streaming networks, which play an important role in network security and network communication. So identifying key nodes in the network is very crucial. The traditional method has huge time overhead for key node identification in largescale networks and cannot guarantee realtime performance. We propose a key node identification algorithm in P2P streaming media networks. Combining with the network structure characteristics of the hybrid mode, we use a regionbased computing model to solve the huge timeconsumption problem caused by the excessive network scale. The importance of nodes is quantitatively described according to the contribution and propagation capacity of nodes. Simulation results show that the proposed algorithm can quickly obtain the results of node importance ranking and effectively identify the key nodes in P2P streaming networks.