The sustainability of social pension insurance is of great significance in guaranteeing the essential life of the elderly and promoting social stability. Based on the provincial panel data from 2012 to 2020, this study uses non-spatial measurement methods, ArcGIS visualization research methods, and geographic detectors to study the regional differences in China’s pension fund balances and the underlying influencing factors. Compared with the traditional way of establishing regression equations to explore the correlation of influencing factors, geographic detectors can quantify the strength of each influencing factor and detect the interaction of different influencing factors. This study found that: First, the growth rate of China’s overall pension fund balances has been declining yearly, with the fastest decline in northeast China, the middle in the Western and Central regions of China, and the slowest decline in Eastern China. Second, the spatial distribution of pension fund balances shows agglomeration characteristics, with high-value areas mainly distributed in Eastern China and low-value regions distributed primarily in Western and Northeastern China. Third, the overall Theil index for pension fund balances is trending down, but the Theil index for the Eastern region is on the rise. Fourth, seven factors, including the working-age population, the population aged 65 and above, and regional GDP, are the main factors that lead to regional differences in the balance of urban and rural residential insurance funds. Finally, the superimposed effects of each element are reflected in double-factor enhancement or non-linear enhancement relation.