In order to improve the identification of transient electromagnetic system echo, a new waveform similarity algorithm is designed to analyze and process the received transient electromagnetic system echo data. Aiming at the fact that the sampled echo data is difficult to distinguish and identify due to various interference factors, and the traditional echo processing method is not ideal, this paper takes the sampling point as an evaluation index from the perspective of space vector, sets the index weight by using entropy weight method, synthesizes various evaluation indexes, and proposes a new waveform similarity algorithm. Through a large number of experimental data analysis, the algorithm has achieved good results in the identification of echoes, and the identification accuracy of partial thickness has reached 90%, which has good practical value.