Target features such as polarization and distance image are often related to many factors, and their values vary with the external circumstances, which greatly increases the complexity and data amount of target recognition. In this paper, a target recognition method based on poles is proposed. Poles are only related to the nature of the target itself, and have nothing to do with factors such as orientation and polarization. However, the real-time extraction of poles is often affected by noise. By extracting poles in an ideal environment, the effect of noise can be reduced and led to a better E-pulse waveform, which greatly reduces the amount of data and complexity required for target recognition. Simulation carried out in Section- III verify the effectiveness and accuracy of the method.