Abstract: Some key nodes in the power grid play an important role in the large-scale cascading failures. In order to improve the identification accuracy of the key node and to prevent the occurrence of the blackout, a new method for identifying the key node based on node traffic importance degree is proposed. The node transmission contribution index and node transmission efficiency are defined by the proposed method, and then the node traffic importance degree of each node is calculated and compared based on the transmission contribution importance matrix of power flow. In the end, the simulation results of IEEE 39-bus system and a southern regional power grid show that the proposed method can quantitatively characterize the importance degree of a node in power flow transmission and electrical topology in the power grid. It has important significance to identify the critical nodes in the network effectively and accurately, and to determine their role in the evolution process of self-organized criticality in AC/DC power grid. It has an important significance on preventing evolution of cascading failure system to the critical state.