Air pollution has been identified as one of the major environmental issues which requires intensive investigation. The detrimental effects of air pollution are well known and well documented. Traffic is considered as a principal source of air pollution in most cities, and it creates an adverse effect on the environment that has an impact on human health globally. Developing a suitable computational model for the accurate prediction of air quality is the need of the hour. The Support Vector Machine based initial prediction model has been attempted to address the problem. The proposed methodology characterizes eight air pollutants concentration (PM10, PM2.5, NO, NOX, NO2, SO2, CO and O3) and four meteorological parameters (RH, WS, WD and Avg. temp.) as model input and AQI as model output. The initial prediction model has been developed using 3 months data (Jan. 2024–March 2024) of monitoring station having peak traffic emission zone at peak hour for Lucknow city. The model achieved encouraging results with high R2 value in case of both training (0.99) and testing (0.96) and low error indices (RMSE value 1.060 and 2.569 for training and testing, respectively).