Aiming at the problem of PM2.5 value prediction in air pollution

takes the hourly monitoring data of Shunyi monitoring station in Beijing as the research object

and a PM2.5 prediction model fuses attention mechanism is proposed. This model effectively utilizes the advantages of long short term memory（LSTM）memorable time series in medium and long term information

and learns the weight distribution between multifeature factors and PM2.5 through local attention mechanism

so as to pay attention to the feature that affect PM2.5 value greatly. improve the final prediction effect. The experimental results show that the evaluation index of PM2.5 prediction model fuses attention mechanism has better performance.