This paper considers the analysis and controller design for the ALSTOM gasifier system. The inherent properties of this highly coupled and numerically ill conditioned multivariable system are studied. A simple reduction is used to determine a minimal realization of the system after using the Osborne transformation to improve the numerical condition of the system. The relative gain array (RGA) is used to achieve suitable input-output (I/O) pairings for the gasifer system, and Edmunds scaling is used to improve the diagonal dominance of the system and numerical conditioning. Model order reduction methods are applied to simplify the subsequent design. A controller is designed using the H-2 methodology at the 100% load condition. The robustness of this controller at other load conditions is assessed and gives satisfactory results. (C) 2003 Elsevier Ltd. All rights reserved.