The generation of realistic 3D microstructure images is crucial for understanding and optimizing materials in various fields, including fuel cell technology. In this study, we propose an approach for generating 3D microstructure images of O₂ fuel cell electrode using Generative Adversarial Networks (GANs) with Minkowski functionals. Our methodology leverages the power of GANs and integrates Minkowski functionals, specifically porosity and specific surface area, as key components during the training process. Minkowski functionals provide valuable geometric measures that capture essential characteristics of the microstructure, enabling a more accurate representation and generation of the intricate details. During training, the Minkowski functionals are incorporated in the loss function to guide the generator’s learning process. By incorporating these functionals into the GAN framework, we aim to capture the complex spatial patterns and statistical properties of the microstructures accurately. Through experimental evaluation, we demonstrate that our approach generates realistic 3D microstructure images of O₂ fuel cell electrode. The implementation code can be found at: https://github.com/a-bentamou/GAN-with-Minkowskifunctionals-for-3D-microstructure-images