Accurately predicting the saturation of the gas hydrate is one of the most critical issues for reserves calculation. At present, empirically geologic and physical constraints dependency equations have been developed to estimate hydrate saturation. However, the models derived from the equations are too simplistic to imitate the complicated submarine-sediments system. We have developed a new method for gas hydrate saturation predictions in the Shenhu area of the South China Sea. This method is different from traditional ones. It is dependent on a statistical estimator of data/samplings. The method solves the inverse problem for which a database consisting of pairs of inputs and corresponding outputs is required. This method uses radial-basis-function (RBF) interpolation to derive a nonlinear mapping function between P-wave velocity increment (database input) and gas hydrate saturation (database output). Such a mapping function is then used for predicting the output when input is known in other places. We defined the velocity increment as the amount of actual velocity deviated from background velocity. The parameters of the interpolating function were calibrated using the database. The test results of this method on well SH2 and well SH7 in the Shenhu area were quite satisfying. The RBF interpolation method is a fast and efficient way for gas hydrate saturation prediction, and it allowed us to bypass complicated physical models’ inference.