In this paper, two new identification schemes based on VSS theory and sliding mode are developed to address the time-varying parameter identification problems. The proposed direct identification scheme can successfully estimate fast time-varying parameters when the process is linear in the parametric space. The proposed indirect identification scheme further exploits the possibility of estimating time-varying parameters for some classes of nonlinear processes not linear in parametric space, but linearizable in the parametric space when the sliding mode exists. Owing to the robustness of variable structure control, processes with fast time-varying parameter uncertainties and nonlinearities are now controllable, and in the sequel identifiable.