A mathematical model that describes the dynamic behaviour of the steam pressure inside a fire-tube boiler has been obtained by using system identification techniques, due to the complexity of the first-principles-modelling of this process. A complete system identification procedure has been carried out, from experimental design to model validation using a laboratory-scale boiler and then ratified in an industrial-scale boiler. In both cases, the identified model is characterized by a second order linear ARMAX structure and time delay which describes with very high precision the steam pressure variation process. This model will be used for model-based control and prediction applications. The reported results show that system identification and model-based control have important roles to play in the management and intelligent use of available energetic resources. © 2008 Elsevier Ltd. All rights reserved.