Abstract: Soil erosion is one of the major threats to ecological protection and high-quality development in the Yellow River Basin, with rainfall erosivity serving as a key driving factor. Quantifying the spatiotemporal variations of future rainfall erosivity is essential for assessing potential soil erosion risks. Based on five CMIP6 precipitation datasets, this study employs the Xie model to project changes in rainfall erosivity over the Yellow River Basin under low-emission (SSP126) and high-emission (SSP585) scenarios for the near future (2021–2060) and the far future (2061–2100), and conducts a quantitative analysis of the contributions of annual erosive rainfall frequency and intensity to changes in rainfall erosivity. The result revealed several critical findings: (1) Rainfall erosivity shows an increasing trend under both emission scenarios, with a more significant rise under the high-emission scenario. Compared to the historical period, rainfall erosivity increases by 9.2% and 12.83% in the near future, and by 14.78% and 30.57% in the far future under the low- and high-emission scenarios, respectively. (2) The projected increase is mainly driven by the rise in both the frequency and intensity of erosive rainfall events, with the average relative contribution of erosive rainfall days exceeding 50%, indicating a dominant role. Seasonal analysis reveals that changes are more pronounced in spring and winter, which warrants further attention. (3) Compared with the high-emission scenario, the low-emission scenario effectively mitigates the increase in rainfall erosivity, avoiding 2.9% and 17.5% of the projected growth in the near and far future, respectively. Adopting low-emission strategies can mitigate the projected increases in annual erosive rainfall frequency and intensity, thereby reducing future soil erosion risks in the Yellow River Basin.