In this note, a discrete-time robust model predictive control (MPC) design approach is proposed to control systems described by linear parameter-varying models in input-output form subject to constraints. To ensure the stability of the closed-loop system, a quadratic terminal cost along with an ellipsoidal terminal constraint are included in the control optimization problem. The MPC design problem is formulated as a linear matrix inequality problem. The proposed MPC scheme is applied on a continuously stirred tank reactor as an illustrative example.