The invention relates to the technical field of power transmission and transformation equipment monitoring, in particular to a weak traveling wave capturing method based on energy change rate kurtosis and phase spectrum order degree, and the method comprises the steps: collecting and processing a traveling wave signal of a line, and obtaining a preprocessing signal; screening based on sliding window kurtosis to obtain an energy mutation event; and judging whether a fault occurs based on the cross-scale phase gradient entropy. In order to solve the problem that online detection and extraction of traveling wave signals are relatively difficult in the prior art, two screening processes of sliding window kurtosis and cross-scale phase gradient entropy based on the energy change rate are introduced, extremely weak signalmutation can be reliably captured from strong noise by detecting the kurtosis of the energy change rate instead of the energy amplitude value, and the signal detection accuracy is improved. The problem of report failure caused by low signal-to-noise ratio of a traditional method is solved. According to the cross-scale phase gradiententropy criterion, fault traveling waves and random noise with physical rules can be effectively distinguished from the internal structure order of the signals, and the false alarm rate is greatly reduced.