Research on soil erosion in agriculture is urgent because it affects soil quality, and agricultural production, and causes environmental pollution. Research and application of science and technology to soil erosion can also contribute to the development of sustainable soil cultivation techniques. Our research aims to address the remaining issues of the RUSLE model [1] in calculating soil erosion levels, errors with actual measurement data, especially the results of erosion and sedimentation calculations, and the complex formulas that make the processing speed slow. We did this by experimenting with different combinations of factors to match the climate, specific properties, and soil composition of the two main landforms in Vietnam, which are mountainous and lowlands. Then, we compared the results with reality to select the optimal combination of factors. The best model is K+C, which produces output results ranging from 13.7 to 674 tons/ha/year and the calculation speed is increased to 1.32 seconds. This is also the desired outcome from an academic stand-point of this research.