Based on the correspondence principles, the analytical visco-elastic solution of round tunnel displacement under hydrostatic pressure is given using Kelvin constitutive law. After simplifying the testing tunnel as a round one subjected to hydrostatic pressure, the preliminary rheological parameters are obtained by using the analytical solution to fit the monitoring displacement data. A FLAC3D model is then established in which the testing tunnel is of the real shape and is under the real ground stress. The preliminary rhelogical parameters are then adopted as parameters required under the creep option of FLAC3D, and the numerical solution of the test tunnel displacement is then given. With the help of BP neural network, the mapping relationship between the rheological parameters and the corresponding numerical solution of displacement are established. And the rheological parameters are then back analyzed by applying monitoring data to this trained network. Finally, the verification of the usefulness and reliability of the back-analyzed rheological parameters is given by implementing a normal calculation process to a tunnel.