Due to the differences in economic development level and fertility policy enforcement, the population age structures are quite diverse from different regions in China. Thirty one provinces and Autonomous regions are used as clustering objects. Population from various age groups and Total dependency ratio are chosen as clustering indexes and whitening weight functions from corresponding indexes are defined. Meanwhile, clustering models of the population age structure assessment are provided and corresponding algorithms are compiled. Thirty one regions are divided into three groups by the age structure: excellent, normal, and poor. Then, sort again within each group to further analyze the reasons. The conclusion from our research provided a systematic, objective and accurate evaluation of age structure in various regions, it may has some reference value for solving China's population problems.