MPSIAC is currently known as an appropriate method to measure sediment of Watershed basins of the country while there has not been any sensitivity analysis so far for this method. In this study, required data for MPSIAC model were gathered from six basins; Amame-Kamarkhani, Kand-Golandok, Tang Kenesht (from two different references), Nojian (from three different references), Pegahe sorkh katvand (from one reference) and Ivanaki (authors’ research). Eleven sensitivity analyses were conducted and the amount of sediment was calculated using the sum of nine factors. Each input parameter was increased or decreased by 20% using a computer program in Visual Basic in Excel. Then sensitivity of the model for the parameters was analyzed. The less sediment has the basin, the less sensitivity has the input parameters in the model. MPSIAC model has twenty input variables that resulted in nine main factors. Erosion parameter (R) was calculated by adding nine main factors and the quantity of sediment was calculated by an exponential function. By evaluation of nine main factors, it was concluded that land use and Gully development were the most sensitive factors. As a result, based on the area and sensitivity of main factors, more investments must be done on the most sensitive factor to reduce soil erosion. In the assessment of nine main factors occurred errors are reduced due to adding operation.Regarding all items, each factor that has more input quantity has the highest sensitivity. As a result, if score of each factor grows more than six, more attention must be paid to score assignation. If occurred errors in assessment of nine factors did not neutralize each other and have additive or decreasing on the R, by addition of 20% to R, it showed that this factor was the most sensitive factor to run the model. If R is equal to 60 and 20% related to the error occurred in calculation of it, 54% error occurs in estimating the amount of sediment and sensitivity reaches to 2.7. This evaluation indicated that MPSIAC model for evaluation of basins with the amount of sediment more than 2.2 ton/hectare must be used more preciously because the model is so sensitive in this status and possible error may get over 50%.