Abstract: The metal magnetic memory detection technology was applied to the detection of pressure bearing components with buried defects. The compression-member were simplified as flat structures, and finite element simulation tests were conducted on flat specimens under uniaxial tension. The influence of different crack sizes and loads on the detection signal was analyzed, and the variation law of magnetic memory signals for weld defects in pressure bearing components was analyzed. The relationship between the characteristic parameters of magnetic memory signals and defects was obtained. Combined with magnetic memory detection experiments, quantitative analysis was conducted on the characteristic parameters and defect parameters of magnetic memory signals, as well as the loading results. The experimental results indicated that magnetic memory signals can effectively characterize defects in compression-members and predict crack propagation trends, providing theoretical support for the application of metal magnetic memory detection technology in compression-member detection.