Air pollution has become an extremely serious problem, with particulate matter having significantly greater impact on human health than other contaminants. The small diameter of fine particulate matter (PM2.5) allows it to penetrate deep into the alveoli as far as the bronchioles, interfering with gas exchange within the lungs. Forecasting air quality has also become important. This study aims to forecast air quality using a combination of multiple neural networks and LSTM to extract spatial-temporal relations. The proposed predictive model considers various meteorology data information related to the elevation space to extract terrain impact on air quality. The model includes trends from multiple locations, extracted from correlations between adjacent locations, and among similar locations in the temporal domain. We also predict the PM2.5 values using regression model in this project. Experiments employing Beijing datasets show that the proposed model achieves excellent performance and outperforms current state-of-the art methods.