In recent decades, losses resulted from natural disasters have dramatically increased and combined with a significant rise in the intensity and frequency of natural hazards, and in exposure. China is prone to multiple hazards and has experienced huge changes in its social fabric during the past two decades, especially in its population density, development patterns, and other social characteristics. These changes may have exposed the population and society to higher risks than ever before. It is, therefore, important to identify and assess regional social vulnerability to natural hazards to strengthen decision making in disaster risk reduction planning and risk governance. To quantify regional social vulnerability to natural hazards and map its temporal–spatial distribution in China, we carefully selected variables from an extensive compilation of provincial socioeconomic and demographic data. Because of the data issue of high dimension and low sample size, we used a two-level approach to calculate social vulnerability index scores. Dynamic geographic patterns of social vulnerability were captured, and spatial autocorrelations were then calculated. The results indicated a quasi-dualistic structure of social vulnerability in China, with a higher distribution in eastern China compared with that in western China, and a declining trend in spatial clustering in recent years.