Abstract: This study conducts empirical comparisons between two mainstream noise models, ISO9613-2 and SPreAD GIS, in Anhui hilly wind farms. By integrating measured, geographic, and meteorological data, the system analyzed the predictive accuracy of the two models under different operating conditions. The results show that due to the integration of terrain, vegetation, and meteorological elements, the overall error of SPreAD GIS is smaller than ISO9613-2, especially when operating at higher power; However, there is still a deviation between the two in low-power and long-range prediction. The research provides a basis for the selection of wind farm noise models, which helps to improve the accuracy of evaluation.