The presented paper shows a new method for modeling of a physical system in a rational expression. The method is based on a hybrid combination of the conventional least square method (LSM) and a robust recursive procedure. In the first step of the method, a sampling function is employed to establish a system of linear equations. Then, the QR decomposition method and a very fast iterative technique are simultaneously used to determine the unknown parameters. Also, it is shown that by regarding a scalar gain parameter, the stability of the final model can be controlled. To verify the performance of the presented method, several theoretical and practical examples are examined and the obtained results are compared with the simulation and measurement data.