Ecosystem services assessment is an important topic in geography and ecology. Current studies on ecosystem services assessment mainly focus on the degradation risk assessment of ecosystem service supply capacity. There are still lacking a risk assessment framework for ecosystem services and related case studies that comprehensively take into account the supply-demand matching characteristics, the changing trends of supply and demand of ecosystem services, and relationships of trade-offs and synergies among ecosystem services. In this paper, we proposed a regional-scale framework for supply-demand risk assessment of ecosystem services by systematical literature review. Then, we applied the framework to water yield service in Shaanxi province to reveal its spatial-temporal patterns of supply-demand risk. Finally, we discussed the influencing factors, significance, and future research directions of supply-demand risk assessment of ecosystem services. The results showed that:(1) in terms of the spatio-temporal pattern of water yield service, the total of water yield supply and demand in Shaanxi province increased from 2000 to 2010. The southern Shaanxi was the main area of water yield service supply, and the regions with massive demand for water yield service mainly distributed in Guanzhong Plain and Hanzhong basin. (2) In terms of supply-demand matching of water yield service, compared with 2000, the areas where water yield supply could not meet the demand in 2010 decreased by 3.93%. Therefore, the supply-demand matching situation improved as a whole. (3) In terms of supply-demand risk of water yield service, the high-risk areas accounted for 13.37% of the whole province from 2000 to 2010, which mainly distributed in Guanzhong Plain and Yulin city. The low-risk regions accounted for 86.83%, which mainly distributed in southern Shaanxi, Yan'an city, and Baoji city. Compared with 2000 to 2005, the supply-demand risk grade of water yield service in Shaanxi province decreased from 2005 to 2010, and the proportion of high-risk areas decreased by 1.63%. The results of the study are expected to provide the theoretical support for the research and management of risk assessment of ecosystem services.