We have studied imprint mechanism of ferroelectric memory based on the oxygen-vacancy screening model. Interface charge density and offset bias for PZT resulted from the imprint were estimated as a function of effective depth for diffusion of oxygen vacancy based on this model. In order to understand microscopic dynamics during the imprint that might affect the interface charge density, molecular dynamics simulation (MD) was carried out. Diffusion length during the real imprint test for PZT was estimated based on the diffusion constant obtained from the MD simulation. By using the diffusion length, interface charge density and, in turn, offset bias was quantitatively estimated. Based on the quantitative analysis, it is found that the oxygen-vacancy screening model with microscopic dynamics obtained from the simulation explains well the imprint mechanism for PZT materials family.