Safety evaluation for medical devices includes the toxicity assessment of chemicals used in device manufacturing, cleansing and/or sterilization that may leach into a patient. According to international standards on biocompatibility assessments (ISO 10993), chemicals that could be released from medical devices should be evaluated for their potential to induce skin sensitization/allergenicity, and one of the commonly used approaches is the guinea pig maximization test (GPMT). However, there is growing trend in regulatory science to move away from costly animal assays to employing New Approach Methodologies including computational methods. Herein, we developed a new computational tool for rapid and accurate prediction of the GPMT outcome that we named PreSS/MD (Predictor of Skin Sensitization for Medical Devices). To enable model development, we (i) collected, curated, and integrated the largest publicly available dataset for GPMT; (ii) succeeded in developing externally predictive (balanced accuracy of 70-74% as evaluated by both 5-fold external cross-validation and testing of novel compounds) Quantitative Structure-Activity Relationships (QSAR) models for GPMT using machine learning algorithms, including Deep Learning; and (iii) developed a publicly accessible web portal integrating PreSS/MD models that enables the prediction of GPMT outcomes for any molecules using. We expect that PreSS/MD will be used by both researchers and regulatory agencies to support safety assessment for medical devices and help replace, reduce or refine the use of animals in toxicity testing. PreSS/MD is freely available at https://pressmd.mml.unc.edu/. Keywords: sensitization, GPMT, QSAR, deep learning,