Understanding the intrinsic characteristics of DFIG is of great importance to reveal the influencing mechanism of DFIGs on power system stability. The conventional analysis of intrinsic characteristics is based on the modal analysis for the detail model of DFIG and characterized by high-order equation and heavy computation. Besides, the results of the analysis lack effective validation. A method of reduced-order model based modes estimation is proposed according to the characteristics of state variables and their coupling property. The multi-time scale and coupling properties of the state variables of DFIG system is analyzed. And then, the reduced-order model based modes estimation of mechanical and rotor electrical oscillatory modes is proposed, and the essence of the method is analyzed using singular perturbation theory. In the end, the correctness of the reduced-order model based modes estimation is verified by comparing the mode results from decomposed-reduced model and detail model.