Diagnostic systems play a major role in the safety of industrial systems and the availability of its equipment. Rapid detecti on (as soon as possible) to the operator of the detected deviations (defects) in relation to the expected nominal behavior is fundamental for the implementation of preventive and corrective actions on industrial systems. These industrial systems are governed by several physical phenomena and various technological components, which is why the Bond Graph tool, based on an energy and multi-physical analysis, is well suited. In this article, we will discuss the problem of diagnosing an active suspension system of a vehicle by presenting diagnostic methods. Then, we introduce the bond graph tool for the robust diagnosis of the system. Finally, the extension of the deterministic models presented to models integrating uncertain element s (Fractional linear transformations LFT) and the generation of robust analytical redundancy relationships are also detailed.