During recent decades in Pakistan, dense urban areas generally underwent more warming than rural areas and small towns. Minimum temperatures also generally increase more than maximum temperatures. This study aims to understand why minimum temperatures increase more in urban areas than maximum temperatures. Simulations are performed with an urban climate model for different configurations of urban parameters: city size, land urban fraction, and building heights. The relative effects of these urban parameters on minimum and maximum temperatures are analyzed, and used to investigate the evolution of temperatures in the megacity Lahore (Pakistan) from 1980 to 2004. Both minimum and maximum temperatures tend to increase with city size and urban fraction. Higher buildings can lead to higher minimum temperature, but also to lower maximum temperature depending on the season. The simple model reproduces well the difference of evolution in minimum and maximum temperature, but overestimates the increase of maximum temperature. Our study indicates that the city size of Lahore appears to be the main factor driving urban warming and its variations.