The magnetic memory method can effectively determine the stress concentration areas of ferromagnetic metal components. However, at present, the magnetic memory signals in the elastic stage and plastic stage of the components are hard to be distinguished, and the stress concentration degree and service life of the components cannot be evaluated effectively. In this paper, the boundary slip model of the magnetic memory effect is built based on the theory of solidstate electronics, and the linear muffintin orbital (LMTO) algorithm is used to calculate the variations of the system energy of the solid and the spin density of states of the electrons at different orbits in the elastic stage and plastic stage. Then, the changing rules of the magnetic memory signals of the components after yielding are quantitatively analyzed. The research results show that the stress concentration degree is in direct linear proportional relationship with the system boundary slip energy and in the inverse linear proportional relationship with the peak to peak value of the electron spin density of states and the magnetic memory signals. After the plastic deformation of the components, the system energy and electronic spin are changed irreversibly, and a turning point appears in the magnetic memory signal curve. The initial value of the magnetic memory signal is getting less and the slope of the curve is getting lower after every plastic deformation of the component.