This paper is concerned with how non-linear physical state space models can be applied to on-line detection of fouling in heat exchangers. The model parameters are estimated by using an extended Kalman filter and measurements of inlet and outlet temperatures and mass flow rates. In contrast to most conventional methods, fouling can be detected when the heat exchanger operates in transient states. Measurements from a clean counterflow heat exchanger are first used to optimize the Kalman filter. Then fouling is considered. The results show that the proposed method is very sensitive, hence well suited for fouling detection. (C) 2006 Elsevier Ltd. All rights reserved.