We investigated a performance of high-temperature proton exchange membrane fuel cells (PEMFCs) with a serpentine flow field through three-dimensional computational fluid dynamics (CFD) simulations and machine-learning-based surrogate models.CFD simulations employed a multiphysics approach using the finite volume method, while deep learning (DL) algorithms facilitated data-driven anal.We aimed to address the high computational cost and data generation challenges in PEMFC simulations by introducing a novel deep transfer learning (DTL) approach that leverages existing data from simpler configurations to efficiently model complex serpentine flow fields.Therefore, given the computationally expensive and resource-intensive simulations of the serpentine flow model, we applied transfer learning (TL) in our DL models, utilizing existing simulation data from straight-channel PEMFC systems.With this framework, we strategically apply existing simulation data from straight-channel PEMFCs to inform the modeling of more complex serpentine configurations.This approach addresses the significant computational challenges associated with simulating serpentine flow fields, enabling efficient and accurate predictions related to complex flow field designs.We developed two DTL algorithms tailored for PEMFC applications, namely, a one-dimensional convolutional neural network (1D-CNN) and a multilayer perceptron (MLP) to predict the c.d. for a given cell voltage.We found that both algorithms exhibit excellent accuracy in predicting c.d., however, the 1D-CNN surpasses the MLP model when evaluated quant. over a wide range of performance criteria.For instance, the R2 values for 1D-CNN and MLP, are observed as 96 % and 91 %, resp., validating their robustness and ultimately the potential of TL in advancing PEMFC technologies.To ensure the explainability of our models, we used SHAP (SHapley Additive exPlanations) technique to explain the decision-making processes of the algorithms.Furthermore, CFD simulations provided detailed insights into the distributions of various parameters within the fuel cells, providing crucial information about the voltage variation impact on performance, thermal management, and high electrochem. regions in PEMFC.