The guaranteed cost control problem for multimodeling systems with norm bounded uncertainty is investigated. The main contribution in this paper is that a new epsilon-independent controller is derived by solving the reduced-order slow and fast algebraic Riccati equations (AREs) whose dimension is smaller than the dimension of full-order multiparameter algebraic Riccati equation (MARE). It is shown that if these AREs have a positive definite stabilizing solution then the closed-loop system is quadratically stable and has the cost bound. (c) 2005 Elsevier Ltd. All rights reserved.