Air pollution is an important international challenge due to its increasing impacts both on public health and the environment. The Air Quality Index simplifies the complex data by transforming the parameters into a single numerical value. However, accurately estimating the AQI stays a difficult because of the complicated connections among the pollutants and weather factors. It is aimed in this study to develop predictive models for AQI using statistical modelling, with a concentration on evaluating the influence of weather variables. The data were gained from publicly available government monitoring stations, specifically the Andhra Pradesh Pollution Control Board, as well as, the Central Pollution Control Board in Visakhapatnam city. Data period was used for the years 2018–2024. The parameters used are: AQI, SO₂, NOx, PM₁₀, PM₂.₅, NH₃, temperature, precipitation, and wind speed. At a particular air quality monitoring station, additional parameters such as NO, NO₂, CO, O₃, benzene, and toluene were included. Linear, polynomial, and exponential models are the models used for the statistical analysis. The results highlighted that the best model with the best effectiveness and fit is the polynomial model, which gives the best performance, the highest R² values (0.889, 0.937, 0.944, 0.900, 0.931, 0.703, and 0.662) across the studied stations were achieved. These findings can support improved air quality management and public health planning in Visakhapatnam city.