Abstract: The value of the soil shear wave velocity V_s measurement value has a very large uncertainty due to the natural heterogeneity of sedimentary soils, operator error, and instrument accuracy. When the measured value of V_s is insufficient, the regression equation between V_s and soil depth h can be established to predict the V_s value under any h. The regression equation is based on the measured value of V_s. A large number of measured data show that the mean and standard deviation of the soil V_s values show significant variations with h. The evaluation of the uncertainty of the V_s values generated by the regression equation in the existing studies mostly focuses on the size of the decidable coefficient of the regression equation and the residual difference between the predicted values and the measured values, and fails to reflect the variation of the standard deviation of the predicted V_s values with depth. Based on a certain number of measured V_s values, the regression equation between soil V_s and h was established by using the SVM (Sediment Velocity Model) proposed by Shi and Domniki, a detailed analysis was conducted on the fluctuation of the uncertainty of the V_s prediction value with respect to h caused by the uncertainty of the fitting parameters in the model through error propagation. Based on the fluctuation, the sensitivity of the fitting parameters to the model was established. The V_s values predicted by the point estimation method were compared and verified using the inverse Nataf transform, and the results were basically the same, proving the reliability of the point estimation method. The seismic response of the site is analyzed for the range of V_s values predicted by the point estimation method, the beta spectra of the mean V_s values predicted based on the point estimation method were found to be significantly larger than the design spectra of the dynamic coefficients for the regulated Class II sites in the period range of 0.25-5 s. Using the V_s value predicted by the point estimation method, it is possible to more reasonably estimate the range of variation of seismic motion parameters in the area.