To study the impact of rainfall spatial resolution on the simulation of urban inundation and the sensitivity of a one-two-dimensional coupled urban inundation model, this research selected the Xiaozhai area of Xi’an City as the study area. Based on three rainfall datasets provided by the Xi’an Meteorological Bureau, the spatial interpolation algorithm of inverse distance weighting was used to obtain five rainfall spatial resolutions: 0.2, 0.5, 1.0, 2.0, and 4.0 km. The Lorenz curve and Gini coefficient were introduced to assess the uniformity of different rainfall spatial resolutions, and they were used to drive the one-two-dimensional coupled urban inundation model for simulation analysis. The results indicate that for the three rainfall events, the Gini coefficients corresponding to the 0.5 km rainfall spatial resolution are 0.163, 0.201, and 0.165, indicating uniformly distributed evaluations. As the rainfall spatial resolution decreases, the Gini coefficient gradually decreases, while the relative errors of peak inundation area, relative changes in inundation volume, and relative changes in inundation depth all gradually increase. In terms of these three hydraulic evaluation indicators, the simulation error is the smallest for the rainfall spatial resolution of 0.5 km, and this resolution is identified as the sensitive rainfall spatial resolution.