Medical image registration is a process of finding the best functional and anatomical mapping between source and target images. Genetic algorithm is the intelligent computation process for solving optimization problems. Medical imaging modalities are available in multiple resolutions which vary for different applications based on diagnosis of a disease. Medical imaging modalities like Magnetic Resonance Images (MRI) are complex modalities showing variations in intensities. These images if registered at one time cannot extract the complete required information. If the same images are registered at different levels of resolution, they can provide more useful information for diagnosis purposes. For example, detecting very small tumors, brain injury detection and computer operated surgery. This paper presents a multi-resolution registration process which reduces mean intensity difference between the source and target images