This work deals with the fault estimation of actuator fault for a polymer electrolyte membrane fuel cell (PEMFC) using descriptor approach. The main contribution of this paper is to consider unknown inputs and time delay in the output. Delay-dependent sufficient conditions for the design of an unknown input observer with time delays are given in terms of linear matrix inequalities (LMIs). For that purpose, an unknown input observer (UIO) for delayed LPV model is designed. Firstly, by considering the faults as auxiliary states, we obtain an augmented LPV system. An algorithm scheme is developed in order to compute the UIO proposed for the augmented system with actuator faults, secondly. Basing on the Lyapunov theory, the convergence and the stability of the UIO are analyzed. Finally and in order to apply the proposed fault diagnosis approach, the time delay is related to the mass flow oxygen in the anode of a PEMFC and actuator faults such as the compressor outlet mass flow rate and the return manifold outlet orifice are designed. Simulation results prove the effectiveness of this approach. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.