The temporal and spatial distribution of air pollution can be analysed with high accuracy using various methods. This study examines the temporal and spatial variation of particulate matter (PM₁₀) and sulphur dioxide (SO₂) pollutants in the Sakarya Basin (Turkiye) using a hybrid modelling approach and presents a spatial hazard analysis. In the study, ground-based measurement data obtained from 16 monitoring stations for the period 2014–2023 were integrated with Sentinel-5 Tropomi satellite data, and calibrationand interpolation methods were applied to produce annual distributions of average and maximum values. Subsequently, the cumulative sum of data from the last ten years was used to reveal the spatial distribution of air quality. The air quality hazard analysis of the basin was modelled using the Random Forest Machine Learning method, employing -ten-year average and maximum result distributions together with topographic wetness index (TWI), land use, precipitation, temperature, road density, wind speed, topographic roughness index (TRI), and -normalized difference vegetation index (NDVI) analysis variables. The study show that PM₁₀ concentrations generally decreased over the last decade, although temporal fluctuations were observed. The highest PM₁₀ values were recorded in 2017 and 2021 and were mainly concentrated around Ankara, Kütahya, Sakarya, and İnegöl. Short-term improvements were observed during the years 2020–2021 when COVID-19 restrictions were in effect. It was determined that SO₂ concentrations entered a significant downward trend after 2015. According to machine learning-based hazard analysis, the areas with the highest risk in terms of PM₁₀ and SO₂ are the city centres of Ankara, Kütahya, Sakarya and İnegöl. The results indicate that air pollution in the Sakarya Basin is critical, particularly around industrial centres and major transport routes, while air quality levels are relatively lower in the northern and southern parts of the basin.