Given the rapid development within China, the inequality of available water resources has been increasingly of interest. Current methods for assessing water stress are inadequate for province‐scale rapid monitoring. A more responsive indicator at a finer scale is needed to understand the distribution of water stress in China. This paper selected Defense Meteorological Satellite Program Operational Line‐scan System night‐time stable lights as a proxy for water stress at the province level in China from 2004 to 2012, as night‐time lights are closely linked with population density, electricity consumption and other social, economic and environmental indicators associated with water stress. The linear regression results showed the intensity of night‐time lights can serve as a predictive tool to assess water stress across provinces with an R2 from 0.797 to 0.854. The model worked especially well in some regions, such as East China, North China and South West China. Nonetheless, confounding factors interfered with the predictive relationship, including population density, level of economic development, natural resource endowment and industrial structures, etc. The model was not greatly improved by building a multi‐variable linear regression including agricultural and industrial indicators. A straightforward predictor of water stress using remotely sensed data was developed.