This paper presents the air quality level by calculating air quality index by using historical data. Air pollution is a pressing environmental issue affecting millions of people in India. Major cities like “Delhi, Mumbai, Kolkata, and Bangalore” frequently experience hazardous air quality levels, posing severe health risks to their inhabitants. Air pollution in Indian cities has emerged as a critical public health concern. Accurate forecasting and analysis of air quality are crucial for implementing timely measures to mitigate these risks. Precise and prompt forecasting of “Air Quality Index (AQI)” is essential for effective mitigation strategies. This research offers a detailed comparison of “Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Random Forest Regression” models to improve AQI forecasting for cities in India. The sequential character of air quality data is versioned by LSTM, which can capture temporal dependencies, Similar functionality is provided by GRU, which streamlines the procedure while preserving excellent temporal pattern recognition performance, while Random Forest makes use of its capacity to manage intricate interactions between meteorological and air quality parameters. The evaluation of the models is carried out with historical AQI data from multiple Indian cities, incorporating “pollutant concentrations such as PM2.5, NO, NO2, NOX, NH3, CO, SO2, O3, Benzene, Toluene, and Xylene.” Experimental results show that integrating “Long Short-Term Memory (LSTM) networks, Gated Recurrent Units (GRUs), and Random Forest Regression” models presents a promising approach to enhancing the accuracy and reliability of “Air Quality Index (AQI) predictions.”