Light pollution is often overlooked, but according to statistics, it has increased by at least 49% globally in the past 25 years. In this paper, the Combination weighting method, metabolic GM (1,1) model, and other methods are used to study the light pollution problem. A light pollution risk level evaluation system is established by using the combination weighting method and several indicators related to light pollution. Based on the analysis of samples from China and the United States, a range of light pollution control strategies is proposed encompassing three key aspects: electricity accessibility, population density, and biodiversity coverage. The light pollution situation of the two locations in the upcoming year is predicted and compared using a combined approach of the metabolic GM (1,1) model, considering various strategies as well as no strategy implementation. Ultimately, it can be seen that the strategy of electricity accessibility is more effective. The establishment of a light pollution evaluation model enables the measurement of the effectiveness of prevention and control strategies, thereby enhancing the ability to effectively manage light pollution.