Deep tissue injuries (DTIs), such as pressure ulcers, result from prolonged mechanical loading on soft tissues, leading to cellular deformation and ischemia, which subsequently cause tissue necrosis and patient morbidity [1]. Conventional Finite Element (FE) models are valuable for predicting strain localization in tissues under mechanical loading but are computationally intensive, limiting their clinical utility [2]. This study presents a dictionary-based Reduced Order Model network (ROM-net) to enable real-time strain predictions in subcutaneous tissues during sitting, overcoming the limitations of traditional FE models. The ROM-net was constructed from FE simulations on a dataset of 16 subjects, further enriched through data augmentation techniques, including sub-modeling and statistical shape modeling. A dictionary of local reduced-order models was created using K-medoid clustering to capture a variety of anatomical and mechanical configurations. A Decision Tree Classifier, trained on the augmented dataset, selects the optimal ROM based on patient-specific characteristics, facilitating rapid, accurate strain predictions. Evaluation on unseen test cases demonstrated accuracy, precision, and F1-score values of 83%, 100%, and 90%, respectively, confirming the model’s robustness. The proposed ROM-net approach significantly reduces computation time from hours to seconds, achieving a projection error of 9% compared to high-fidelity FE simulations. This framework offers potential for real-time strain assessment in clinical settings, providing a proactive tool for pressure ulcer prevention. Ongoing validation efforts aim to facilitate its integration into clinical workflows. [1] C. W. J. Oomens, “A Multilevel Finite Element Approach to Study Pressure Ulcer Aetiology,” in Multiscale Computer Modeling in Biomechanics and Biomedical Engineering, in Studies in Mechanobiology, Tissue Engineering and Biomaterials. Springer, Berlin, Heidelberg, 2013, pp. 289–298. doi: 10.1007/8415_2012_158. [2] A. Macron, H. Pillet, J. Doridam, A. Verney, and P.-Y. Rohan, “Development and evaluation of a new methodology for the fast generation of patient-specific Finite Element models of the buttock for sitting-acquired deep tissue injury prevention,” J. Biomech., vol. 79, pp. 173–180, Oct. 2018, doi: 10.1016/j.jbiomech.2018.08.001.