Off-line dynamic identification requires the use of a model linear in relation to the robot dynamic parameters and the use of linear least squares technique to calculate the parameters. Most of time, the used model is the Inverse Dynamic Identification Model (IDIM). However, the computation of its symbolic expressions is extremely tedious. In order to simplify the procedure, the use of the Power Identification Model (PIM), which is dramatically simpler to obtain and that contains exactly the same dynamic parameters as the IDIM, was previously proposed. However, even if the identification of the PIM parameters for a 2 degrees-of-freedom (DOF) planar serial robot was successful, its fails to work for 6 DOF industrial robots. This paper discloses the reasons of this failure and presents a methodology for the identification of the robot dynamic parameters using the PIM. The method is experimentally validated on an industrial 6 DOF Stäubli TX-40 robot.