This paper presents an elaborate scheme, which is different from the traditional parameter identification method, to estimate the parameters precisely for high-speed magnetic suspension motor. It is well known that the inductance and resistance of each winding play an important role in ensuring the stability of the controller for motor. There have been sufficient research works on equivalent circuit model, where the input and output data are assumed to be true value, and unknown circuit parameters are obtained by solving the equivalent circuit equations of motor. However, the output data are often interfered with noises, such as measurement error, the high-frequency components from pulse width modulation signals and sampling errors. The system response output data are reconstructed through applying in stochastic theory, which is different from the traditional filter techniques, to eliminate effectively the above-mentioned noises. Thus, the parameters can be estimated precisely. As an experimental verification, parameters estimated by the proposed method are applied in calculation of switching point between pulse width modulation mode and pulse amplitude modulation mode for magnetic suspension motor. The experimental results validate the effectiveness of the proposed method.