<p>The aim is to predict extreme hydrological events under the influence of climate change and land use and land cover (LULC) by 2050. In this study, the modified SCS-CN hydrological model was applied to the Amazon Basin. The innovation is to provide a robust tool to quantify future extremes events, using Google Earth Engine (GEE) for the dynamic parameterization of curve number (<i>CN</i>) and initial abstraction (<i>I</i>_<i>a</i>), thereby overcoming empirical limitations. Climate uncertainty was mitigated by regionalized selection of the best GCMs for each subbasin. The model was calibrated and validated with runoff from the state-of-the-art global reanalysis dataset for terrestrial applications (ERA5-Land), demonstrating high accuracy (Kling–Gupta efficiency (KGE): 0.62 to 0.93; Pbias: 0.4% to 8.0%). The standardized runoff index (SRI) was used as the indicator of extremes. The model showed sensitivity, which was proven by capturing the historical drought of 2023. The methodology implemented in GEE has made it possible to analyze surface runoff in the future. The findings are: (1) In the Alto Tocantins and Araguaia subbasins, the increase in climate forcing does not translate to the intensification of droughts, suggesting that these subbasins have reached a hydrological tipping point. (2) The projections indicate a drier Amazon, with an increase in the frequency and duration of droughts during the dry season. (3) The increase in the frequency of extreme droughts, aggravated by LULC, represents a direct threat to water security. Therefore, the developed interactive application constitutes a useful public planning tool applicable to other river basins.</p>