This paper applies spatial autocorrelation statistics to analyze the spatial ifferentiation and transition of aging population during the period of 1980 and 2000 based on data col1ected from 106 administrative units in Central Taiwan. Moran’s I statistic is employed in this research to measure the spatial association of population aging. Results have shown that the distribution of aging exhibits a spatial concentration and expansion over time. In addition, spatial autocorrelation correlograms based on Z(I) ,statistic have also shown that the distance among spatial clusters tends to decrease during the study period. Finally, this study has identified the actual clusters according to local index of spatial autocorrelation. It suggests that most clusters of aging in the study area appear mainly in mountain or rural areas where slower development have occured, while clusters of younger population, on the other hand, concentrate mostly in urban areas. The fact that the number of clusters of aging population increases and the value of Z(I) statistic decreases throughout the study period suggests that distribution of aging population becomes less ,concentrated over space. It can be concluded that as population aging in Taiwan becomes more common in recent years, the aging population tends to be more evenly distributed over space. This finding is quite different from conclusions drawn in earlier studies. The distribution of the younger population, however, has shown a tendency towards spatial concentration within a few relatively isolated clusters.