The research on SO pollution in China has been hotly debated over the past decades. Different from the existing studies, this work employs satellite observed SO columns from 2005 to 2016 and applies a spatial econometric approach to investigate the socio-economic influencing factors of SO pollution of 270 prefecture-level cities in China. The findings are as follows. (1) SO pollution over China exhibits a significant and positive spatial autocorrelation. (2) The most polluted area is concentrated on the North China Plain. However, SO pollution over China has been reduced gradually during the sample period, implying that overall environmental quality in China has been substantially improved. (3) Besides, the results of spatial econometric models are not in support of "pollution haven hypothesis". On the contrary, the pollution halo effect of foreign direct investment works well and contributes to reducing SO pollution in China. Moreover, we find that urban economic levels and innovative capability are negatively correlated with SO pollution, indicating that economic growth and an increase in innovation can help improve environmental quality. On contrast, the share of the secondary industry, urbanization and transportation are found to have positive impacts, indicating that they are three main contributors to SO pollution in China.