In steam power-plants, to prevent over-heating of drum components or flooding of steam lines, perfect control of drum water level is of great importance. But during the operation, disturbances affecting water level, model uncertainties and parameter mismatch due to variant operating conditions lead to the variation of model parameters. In this paper, under transient conditions and in the presence of model uncertainties, two control strategies are implemented to achieve desired tracking of drum water level: robust sliding mode and H-infinity control. Two transfer functions between drum water level (output variable); feed-water and steam mass rates (input variables) are considered. For the dynamic system with time varying characteristic and parametric uncertainties, a sliding mode controller is developed and an optimal H-infinity, controller is designed based on mu-synthesis with DK-iteration algorithm. For different desired commands of drum water level (including a sequence of steps and ramps-steps); it is observed that both control strategies guarantee robust stability and performance of the system without actuators saturation (control signals are bounded). However, using the sliding mode controller leads to the more smooth and rapid time responses of drum water level with less oscillatory behaviour of control efforts (and consequently less energy consumption). In addition, for tracking objectives in short command times, sliding mode controller performs more appropriately. (C) 2012 Elsevier Ltd. All rights reserved.