In order to reduce power information communication errors and realize accurate dispatching, it is of great significance to accurately identify abnormal signals of distribution network carrier communication. Therefore, an abnormal signal identification method of distribution network carrier communication based on parameter estimation is proposed. This method defines the distribution of communication nodes by establishing the distribution network carrier communication network model. Then, taking the constructed distribution network carrier communication network model as a reference, the collector is deployed on each node to collect the carrier communication signal passing through the node, and the independent component analysis method is utilized to denoise. The carrier communication signal parameters of distribution network are estimated by the combination of genetic algorithm and fractional Fourier transform. Finally, the two estimated values are composed of signal features, which are used as input to realize the abnormal signal identification of distribution network carrier communication through classifier. The results show that compared with the traditional identification methods, the proposed method can identify the abnormal signal of distribution network carrier communication more accurately, and has higher sensitivity and specificity.