A limited number of atlases have been constructed using quantitative susceptibility mapping (QSM) images from subjects with Parkinson’s disease (PD), and given disease-specific subcortical structures. In this work, we generated three standard-space templates i.e. the hybrid, QSM and T1w atlas, which kept good image quality to observe brain white, gray matter, and deep-brain nuclei. Based on the atlases, we achieved the manual annotation of a few brain subcortical structures, e.g. globus pallidus, substantia nigra, subthalamic nucleus and thalamus. The results gave the position and shape of subcortical nuclei which could be meaningful for the research and surgical treatment of PD.