The commonly used data reduction methods to determine shear wave velocity (Vs) require the depth interval of the Vs profile to be the same as the in-situ testing interval. The Vs values derived in this way are the mean values within the depth intervals. However, the in-situ multilayer interfaces do not always coincide with the predetermined testing depths, and the materials may change dramatically across layers and thus the above interpretation methods may not always produce representative soil properties. In this study, a modified refracted ray path method for more flexible Vs calculation is proposed. The method takes advantage of the routine cone penetration testing data obtained in SCPT tests and applies a combined approach for the soil layer identification using both the corrected cone tip resistance (qt) and the Soil Behavior Type Index (Ic) as indicators. Vs is then calculated using the Snell's law based on the original shear wave travel times and the newly identified layer profiles. The proposed method can avoid the error caused by crossing the multilayer boundaries and provide a more accurate prediction of soil type and the corresponding stiffness properties. The data show that by using the proposed method, the accuracy of the estimated Vs can be improved by up to 77% for a soil layer with interface not located at testing depths, as compared to Vs obtained from the conventional methods.