Abstract: As the global energy system transitions toward low-carbon and clean energy, utilizing hydrogen for power generation and producing and storing hydrogen through water electrolysis has gradually become a key approach to reducing dependence on fossil fuels and accelerating the transformation of power grids toward new energy. Hydrogen production and power generation units are key components in this system, with their performance, stability, and lifespan directly impacting the overall efficiency and application prospects of the hydrogen energy industry chain. Proton exchange membrane fuel cell power generation and water electrolysis hydrogen production technologies hold promising prospects, but cost and durability remain areas requiring improvement, limiting their further commercialization. First, it briefly describes the advantages of hydrogen production and power generation based on proton exchange membrane technology compared to other methods. Subsequently, the failure mechanisms and diagnostic methods for proton exchange membrane electrolysis cells are elaborated. For typical failures such as membrane degradation and flow field blockage, the causes of different failures are analyzed. The impact on critical components within equipment and corresponding mitigation measures are addressed. Model-based and data-driven fault diagnosis methods are reviewed, comparing the principles, research outcomes, and advantages and disadvantages of different diagnostic approaches. Then, the working principle of proton exchange membrane fuel cells was outlined, and the mechanisms behind typical flooded and membrane dry failures were analyzed, along with their impact on equipment. The latest advances in fault diagnosis methods for proton exchange membrane fuel cells are reviewed, comparing the principles, research outcomes, and advantages and disadvantages of test-based, model-based, and data-based diagnostic approaches. Finally, future development directions are proposed across four dimensions: fault mechanism and model establishment, online diagnosis and multi-fault classification, fault mitigation and control strategy optimization, and integrated advancement. These include: constructing multi-physics mechanism models for diagnostics that incorporate aging patterns; developing online diagnostic methods adapted to actual operating conditions to enhance composite fault recognition capabilities; optimizing control and mitigation strategies to achieve adaptive regulation; Leveraging established fuel cell expertise to develop diagnostic systems tailored to electrolyzer characteristics, thereby enhancing overall system reliability.