In many ways, our everyday lives depend on having access to reliable data about the state of the air around us. If you can predict the air quality ahead of time, you can put in place the right warnings and safety measures to keep people from getting hurt. The Control Boards in India set up the National Air Monitoring Programme (NAMP), which checks the air in 342 locations in 240 cities. There are a few distinct categories for the Air Quality Index (AQI). The AQI in Chennai was predicted using data that was collected and preprocessed to account for missing values and eliminate duplicates. Air quality forecasting using deep learning technology is investigated using a huge dataset describing the surrounding environment. This study suggests a scheme for classifying AQI values using multi-output and (MMS) based on long short-term memory (LSTM). Increased classification precision is achieved by using Chaotic Hunger Games Search (CHGS) in the hyper-parameter tuning process. When compared to conventional methods, the AQI value provided by the proposed deep learning model is more precise and accurate for a given location within a metropolis. The suggested deep learning algorithm improves forecast accuracy, serving as a public service announcement to bring levels down to a safe level. The AQI values can be reliably predicted by the deep learning method, which aids in sustainable urban development planning. By coordinating road traffic signals and encouraging people to take public transit in strategic areas, the AQI target value can lessen pollution.