Determination of fast shear wave velocity by methods such as core analysis requires considerable time and cost, and also are not available at all distances and wells. It is practically impossible to core in horizontal wells, so the use of indirect methods that can be used to estimate shear wave velocity is suitable. In this study, the artificial post-diffusion feed network and the Levenberg-Marquardt training algorithm was designed using a part of the data; then using this artificial network and the remaining data, the correlation of predicted and measured values for fast shear wave velocities was Estimated. Due to the values of correlation and error, it is possible to conclude that this method is a reliable model for predicting the velocity of shear wave velocity and can be used to predict the shear wave velocity of wells without shear wave velocity data.