Abstract: A defect identification and classification method for 444 ferromagnetic stainless steel welded pipes was proposed based on eddy current testing technology. In this method, the eddy current signal of the steel pipe after magnetic saturation processing was first collected, and the eddy current signal was further denoised by empirical mode decomposition (EMD), and the intrinsic modal component (IMF) which can best represent the original signal characteristics was extracted, and the time-frequency domain characteristic parameters were extracted from the selected IMF. In order to improve the efficiency of model recognition, principal component analysis (PCA) was used to reduce the dimension of eigenvector set. Finally, support vector machine (SVM) was used to identify and classify weld defects. The experimental results showed that the method had high accuracy in identifying the defects of 444 stainless steel welded pipe, and can be used to select the products with specific defects automatically to improve the production efficiency.