A biomechanical simulation method for co-registration of breast MRI and low-dose chest computed tomography (LDCT) images is presented, by aligning the images in a virtually unloaded configuration. The breast tissue was considered as neo-Hookean material, and the finite element method was applied to simulate the deformation from gravity-unloading. The Demon’s non-rigid registration algorithm was applied to co-register the gravity-unloaded MRI and LDCT models. Fourteen normal subjects who received both breast MRI and LDCT for breast and lung cancer screening were analyzed. The results show that the pre-processing using gravity unloading can facilitate the co-registration of LDCT and MRI.