Beijing’s rapid urbanisation has driven economic growth and technological advancement while intensifying infrastructure challenges, air quality issues, and temperature rise. Utilising advanced machine learning (ML) and numerical physical modelling techniques, this study quantitively assesses the impact of the future (2030s) urbanisation on surface air temperatures in Beijing, a fast-growing Chinese city, by comparing these effects with those induced by past urbanisation (1990s and 2010s). The boosted regression tree model is used to predict changes in land use. In contrast, the Weather Research and Forecasting (WRF) model with 2-km horizontal resolution is applied to simulate the regional climate of January, July and September from 2008 to 2012. The control simulation results (2010s) are validated against observations, showing that the WRF model reasonably reproduces diurnal surface air temperature variations. Seasonal urbanisation impacts on surface air temperatures reveal a ∼1.0 °C increase from the 1990s to 2010s and ∼2.0 °C from the 1990s to 2030s in the city’s central area. Additionally, the rapid expansion of the low-urbanised regions in Beijing leads to temperature increases of 1.1 °C (1990s to 2030s) and 0.8 °C (2010s to 2030s) at 0500 LST in July, as urbanised land grows from 384 km² (1990s) to 1764 km² (2030s).