Abstract: Ultrasonic detection was utilized to receive signal from corrugated pipe model and the EEMD method (ensemble empirical mode decomposition) was used to decompose signal into intrinsic mode function of multiple frequency spectrums (the IMF). When there is a defect in corrugated pipe, the echo signal's energy distribution in different frequency spectrums will be different. The main IMF component's energy entropy was taken as the input vector of SVM (support vector machine), then the mechanism of classification was set up to justify corrugated pipe's structure. The experimental results show that the proposed method can effectively judge the corrugated pipe's quality.