An algorithm of parameter identification by model compensation auto disturbance rejection controller is given. It is very efficient to the parameter identification problems of low order systems. At the same time it is not easy to get satisfying parameters of the extended state observer when the order of the controlled system is high. To solve the above problem, we combine in this paper the auto disturbance rejection controller constructed by cascade extended state observer(CESO) with the algorithm of successive approximation. Simulations show it can identify the parameter of systems with known structure effectively. In the meantime, the parameters of CESO are much easier to adjust than those of the former ESO when it is applied to high order systems.