This paper proposes a robust control approach for IPMSG based direct drive wind energy conversion systems to extract maximum possible power under varying weather condition. The proposed control approach is robust against parametric and modelling uncertainties. The proposed approach includes nonlinear back-stepping control with sliding mode observer (SMO) theory. The SMO is used to estimate the disturbance in the IPMSG based wind energy conversion system model arising due to parametric and modelling uncertainties. The effectiveness of the proposed approach has been validated through rigorous simulation studies carried out on a grid connected direct drive IPMSG wind energy conversion system.