A new robust model predictive control (MPC) technique is proposed for norm-bounded uncertain systems with input constraints. In order to improve feasibility and system performance, we propose an LMI condition for the cost monotonicity by using a new parameter dependent terminal weighting matrix. We formulate the problem as a minimization of the upper bound of infinite horizon cost function subject to the LMI condition for the cost monotonicity. A numerical example shows the effectiveness of the proposed method.