The virtual propeller model can achieve the rapid numerical prediction of the ship self-propulsion performance through viscous flow, which used the improved body-force method. The two-dimensional lift coefficient CL and the drag coefficient CD are very important parameters in this method, which are generally obtained by the potential flow methods and cannot incorporate viscous effects. This study will perform a fully nonlinear unsteady RANS (Reynolds Average Navier-Stokes) simulation to get the KP505 open-water characteristics and then divide its blade into several parts to get the lift coefficient CL and the drag coefficient CD on each one. Then fitting by multivariate regression method, the relationship between CL, CD and propeller parameters is obtained. The Unsteady Blade Element Theory (UBET) is coupled with RANS in house CFD code HUST-Ship (Hydrodynamic Unsteady Simulation Technology for Ship) to calculate the flow around the propeller. RANS equations are solved by the finite difference method and PISO arithmetic. have been made using structured grid with overset technology. The results show that comparing with the EFD data, the maximum differences of the result of the improved body-force method are 4.32% and 2.7% for the thrust coefficient and the torque coefficient respectively near the propeller operating point.