With the continuous development of society, light pollution has become a hot issue in the field of regional governance and sustainable development. It is of high research value to make accurate and based classification of light pollution risk levels in a region. In this paper, 31 provinces and regions in China were selected as research samples, combined with social development, population, and other five levels, and multi-stage indicators were used to construct a light pollution risk level evaluation index system. The subjective and objective weights of each index were determined by the analytic hierarchy process (AHP) and entropy weight method, and the combined weights were determined by minimizing the deviation between the basic weights based on game theory. The results showed that the top three weights were regional per capita GDP, urban population density, and per capita consumption expenditure of all residents respectively. On this basis, a rank-sum ratio comprehensive evaluation model was established. The light pollution risk levels of 31 research samples were divided into four levels, including 3 high-risk areas, 13 medium-high-risk areas, 13 medium-risk areas, and 2 low-risk areas.