The device under consideration drives a cabin pressure control valve of modern passenger aircrafts. It is already provided by a wide range of built-in tests. However, these tests do not allow a deep diagnosis of the actuator. Therefore in this paper a model-based approach for detecting and isolating faults is presented. Parameter estimation and parity equations are used for feature generation. The features are compared to that one of the fault-free case. Deviations are fed into a fuzzy rule base in order to isolate the faults. The whole approach is implemented on a 16 bit microcontroller. Experimental results are given.