Soil erosion ranks amongst the major causes of land degradation affecting agricultural productivity, soil fertility, water availability, and ecological sustainability. This paper attempts to map soil erosion intensity in the Gundlakamma watershed, which includes Bapatla, Palnadu and Guntur (three districts) in Andhra Pradesh, by using the Revised Universal Soil Loss Equation (RUSLE) coupled with Google Earth Engine (GEE). They compiled multi-source geospatial datasets like GPM rainfall data, Cartosat DEM for elevation, Sentinel-2 satellite images from Earth observation satellite of the European Space Agency, NBSS&LUP soil map (soil information) to arrive at factors viz. rainfall erosivity (R), soil erosion susceptibility (K), length and steepness of slope (LS), management of surface cover (C), and land conservation factors (P). These parameters were combined in a raster-based spatial analysis to estimate overall annual soil erosion and pinpoint erosion-prone areas across a total area of 5959 sq km. It was found that the rainfall erosivity varied from 414.19 to 581.81 MJ mm ha⁻¹ h⁻¹ yr⁻¹, whereas the values of LS revealed the strong topographic control over soil erosion processes. As a result, the basin was further grouped into five levels: Mild (6.43%), Moderate (9.64%), Major (24.99%), Majorly Severe (38.11%) and Very Severe Erosion (20.82%). Nearly 83% of the basin is at high to very critical risk of erosion, which emphasises the potential for significant soil loss. The model performance was checked by Frequency Ratio (FR) validation. The highest FR was assigned to the class of very severe erosion (1.079), and the model overall achieved an FR of 1.400. The validation results support that the RUSLE-GEE method is highly reliable. The paper reveals locations of soil erosion hotspots, where immediate soil conservation is necessary, and the potential of online geospatial applications for the assessment of soil erosion at the watershed level and the formulation of strategies for land use planning.