Abstract: In order to reduce the false alarm rate and calculation cost in the application of smart meter, a measurement infrastructure attack identification method based on time fault propagation graph support vector machine is proposed. The support vector machine was established and trained to detect suspicious behavior in smart meter. The attack path was generated by using the time fault propagation graph technology to identify the attack events. The proposed pattern recognition algorithm was used to calculate the similarity between the detected abnormal events and predefined network attacks. The simulation results on AMI test platform show the effectiveness of the proposed method.