In this paper, a linear quadratic regulator (LQR) is designed to control a nonlinear model of an electronic throttle. First, the model is linearized in the form of a linear parameter-varying (LPV) state-space system using the mathematical equations. This model gives insight about the elements to be used in the regression vector, whereby the second step is to identify the system using a linear regressor by minimizing a least squares criterion. The scheduling parameters in both cases are the angle of the throttle and its speed. Finally, the LQR is designed separately for each one of the identified LPV systems. To cope with the input disturbance existing in the parameter varying models of the throttle, the LQR is implemented with disturbance feedforward. And to cope with the reference tracking problem, either a reference feedforward or an integral effect on the output error is added. This gives rise to 2 different control architectures which will be presented and compared in this paper. The control is later validated on the initial nonlinear model.