Voxel-based morphometry (VBM) is an automated method allowing identification of anatomical differences in the whole brain without the pre-specification of a region of interests. Spatial normalization is one of the major processes in VBM, which transforms all images to a standard template. A variety of templates were employed in VBM researches in the literature, including the MNI template and study-specific templates. Few studies were presented to test the effect of templates on the detection accuracy of VBM, although it is claimed in many papers that the study-specific template performs better. However, the creation of the study-specific template differs on the subjects included. In this paper, the gray matter (GM) difference of two groups (female vs male) was analyzed to evaluate the effects of templates on the VBM results. The statistic parametric mapping (SPM) package, as the standard software for VBM implementation, was used for analyses.