A computationally simple approach is described for obtaining a mathematical model representing the behavior of nonlinear systems over the entire range to which it is subjected. This model is derived using either experimental results or computer simulation of the system under consideration. It is shown that if the information about the nonlinear system is based on experimental results, then the derived model is not affected with the measurement noise.