Abstract:To solve the problem about the evaluation of forest air,an eveluation model based on BP neural network was studied.According to the characteristics of forest air with low pollution and high-volume negative ions,the model was constructed by MATLAB with the indices of SO2,NO2,PM10,O3 and negative ion.The evaluation results are 9.13 and 8.56 respectively by comparing the statistic data of air quality in forest with those in the town in Yichun,which proves the feasibility of the model.