Air pollution is a major problem in many countries and especially in India. In November 2019, New Delhi recorded an air quality index level of 900, which is considered higher than the ‘severe’ level. The air pollution forecasting method predicts the pollution based on the available dataset and the data features and certain method performance in forecasting at high accuracy depends on the method and the measures used. Forecasting air quality levels in countries like India is very important because it has a direct impact on public health, and thus, is used for decision making. The main goal for this paper is to investigate air quality index prediction based on different algorithms, so experts can identify the methods that require development and it is useful as a starting point for novice researchers. The problem of the Air Quality Index (AQI) prediction in this paper is approached with different Fuzzy Inference Systems (FIS), Neural Networks, Swarm intelligence techniques and so on. The results of the experiments shows that the LSTM model performs better than all the deep learning based models and Fuzzy based models discussed in this paper