Abstract: To identify the different damage types of plywood, a method based on timefrequency and proportion of energy was proposed for feature extraction of acoustic emission (AE) signal of plywood by combining the waveletpacket timefrequency analysis with energy spectrum. The result showed that dilatational wave and flexural wave were the main modes of fracture signal from threeplywood outer with wide frequency spectrum, and the energy of signal was mainly concentrated in the first, second, third, fourth and seventhband of the wavelet power spectrum. Rupture signals from table and the entire board of five plywood were mainly dominated by dilatational wave and flexural wave mode respectively, and the former frequency was unitary and amplitude was higher, the latter energy mostly focused on the first, second band. Degumming and the entire board fault of three boards signals waveform were composed of dilatational wave and flexural wave and the latter was dominant, and signals energy mostly focused on the first, second, third and fourth band of the wavelet power spectrum. The test results revealed associative analysis of frequency spectrum, timefrequency map and wavelet packet energy spectrum could be used to identify AE signals characteristics corresponding to various loading failure modes respectively.