The quantitative analysis of soil erosion changes over 7 years due to mining operations in two neighboring hilltops in West-Singhbhum District, Jharkhand, are reported. CartoSat-1, ETM+ and LISS-IV satellites' data provided spatial inputs in Universal Soil Loss Equation (USLE) and Morgan, Morgan and Finney method (MMF) models, which were used to predict the average annual soil erosion during the period of 2001–2008 in a geographic information system (GIS), in six distinct classes. In the comparative analysis of the 7-year period, the MMF model revealed a lower coefficient of variation 0.71 (2001) and 0.84 (2008) in predicted average annual soil loss, which increased by 16% (81.3–94.2 t ha−1yr−1), whereas in the case of USLE, the coefficients of variation were 3.88 (2001) and 1.94 (2008), with an increase of 61% (48.56–78.38 t ha−1yr−1). The correlation coefficient of these models was 0.1 (2001) and 0.36 (2008), which shows that both models predicted significantly differently as a result of the different factors considered. Overall, the MMF model predicted a higher soil erosion rate but less variation than USLE. Both models showed soil erosion rates were drastically increased by anthropogenic activities in the area, hence careful consideration is needed. The same sensor and imaging data could not be maintained. Correction of errors may reduce erosion, but it will still remain significant for future planning.