Basing on a physical and operational analysis of a power grid， a method to identify the power grid's topological model by mining the networks data based on complex networks theory is proposed. Firstly， a directed-weighted topological model based on active and reactive network decomposition is proposed. Several definitions， algorithms and new parameters including the shortest electrical path， the direction and weight of line and the average contributiveness are given at this step. Then data mining method is applied to the topological model. Results of a numerical example show that it has a better reflection to the characteristics of power grid based on the directed-weighted topological model under the active and reactive network decompositions. Meanwhile， it has been verified that the model proposed is rational; and it can effectively identify the important part and area of a power grid based on complex network theory. The results can be used as a reference to the power system security and reactive power control partitioning analysis.