To achieve a good performance of the utility boiler, dynamic variables such as drum pressure, steam temperature and water level of drum must be controlled. In this paper, a linear time invariant (LTI) model of a boiler system is considered in which the input variables are feed-water and fuel mass rates. Due to the inaccessibility of some state variables of boiler system, a minimum-order observer is designed based on Luenberger's model to gain an estimate state of the true state x. Low cost of design and high accuracy of states estimation are the main advantages of the minimum-order observer; in comparison with previous designed full-order observers. By applying the observer on the closed-loop system, a regulator system is designed. Using an optimal functional code developed in MATLAB environment, desired observer poles are found such that suitable time response specifications of the boiler system are achieved and the gain and phase margin values are adjusted in an acceptable range. However, the real dynamic model may associate with parametric uncertainties. In that case, optimum region of poles of observer-based controller are found such that the robust performance of the boiler system against model uncertainties is guaranteed. (C) 2010 Elsevier Ltd. All rights reserved.