The authors present a sliding mode-based adaptive input-output linearising control for induction motors, which covers the more realistic situation in which the load torque and the rotor resistance are not exactly known. A nonlinear identification scheme based on the continuous Hopfield network is proposed which models the unknown parameter deviations between load torque and rotor resistance and their nominal values. Once these are identified, the motor flux amplitude and speed are separately controlled by sliding mode control with variable switching gains. Simulation results show the robust behaviour of the proposed controller in the presence of parameter uncertainties and disturbances.