In the paper an application of the interval fuzzy model (INFUMO) in fault detection for nonlinear systems with uncertain interval-type parameters is presented. A confidence band for the input–output data, obtained in the normal operating conditions of a system, is approximated using a fuzzy model with interval parameters. The approximation is based on linear programming using l_∞-norm as a measure of the modelling error. Applying low-pass filtering when obtaining the confidence band makes it possible to use arbitrary sets of identification input signals. An application of the INFUMO in a fault-detection system for a two-tank system is presented to demonstrate the benefits of the proposed method.