Poor water management usually leads to various degrees of flooding and exacerbates oxygen starvation, both of which affect the performance and durability of a proton exchange membrane fuel cell (PEMFC). This paper proposes a model-based approach to estimating the liquid saturation and current density difference simultaneously. The current density difference is a parameter that indicates the oxygen starvation level. For estimation purposes, a fuel cell cathode model with separate inlet and outlet subsystems is developed to incorporate the effect of harmful phenomena such as flooding and oxygen starvation on the system dynamics and cell voltage. The cathode gas diffusion layer (GDL) flooding and oxygen starvation diagnoses are formulated as state estimation problems. The proposed approach is validated through an offline simulation using experimental data. The offline prediction suggests the effect of the anode purge and air steam conditions on the internal states of the PEMFC, which is helpful for system optimization and control design. The estimation problems are further decoupled, and a simplified algorithm is designed for onboard applications. Finally, the modified fuel cell model and optimized estimation algorithm are applied to an online test system as a demonstration. Copyright (C) 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.