Matching pursuit method based on differential evolution (DE) Algorithm is presented to process ultrasonic guided-waves testing signals. The Gaussian modulated functions is chosen as matching atoms, whose time-frequency characteristics agree well with that of ultrasonic guided-wave signals. DE Algorithm is employed to improve parameters searching efficiency, and cross0-term free time-frequency distribution (TFD) is achieved by superimposing the Wigner-Ｖille distribution(WVD) of each matching atom decomposed. The processed results of the testing signals show that, matching pursuit based on DE algorithm can greatly improve parameters searching efficiency, and the defects discrimination can be achieved from echo signals by TFD characteristic comparison.