Aiming at the slow convergence speed and low accuracy of traditional air quality prediction models,a predication model based on variational mode decomposition(VMD)and dung beetle optimizer(DBO)was proposed to optimize long short term memory(LSTM).First of all,for the problem that the AQI raw data has a large amount of noise,the VMD method was used to decompose the nonstationary data to reduce the influence of noise on the prediction results,so as to obtain multiple modal components with different features.Secondly,there are some limitations in rely on manual parameter tuning based on human experience for LSTM,the DBO algorithm was used to optimize the LSTM model parameters.Finally,the LSTM model was used to predict each subseries after decomposition,and the subseries are superimposed to obtain the final prediction result.The experimental results show that the decomposition of nonstationary data by VMD can help improve the prediction accuracy,and the performance of VMD-DBO-LSTM model is improved to varying degrees compared with other models,the root mean square error of this model prediction is 4.73μg/m,the average absolute error is 3.61 μg/m³,the goodness of fit reach 97.8%.