We find a control scheme for a squirrel-cage induction motor, where the aim is to maximize the power factor and linearize the total electromagnetic torque supplied by the motor. Thanks to this maximization of the power factor, we represent the nonlinearities of the motor model in an algebraic equation; and thus we linearize the model by a simple change of variable. Since this linearization is very sensitive to parameter uncertainty, we develop an adaptive control law that compensates such uncertainties in order to guarantee accurate linearization.