Blood sample collection and laboratory testing are necessary for traditional blood group detection, and these procedures can be time-consuming, intrusive, and resource-intensive. This paper presents a non-invasive technique for classifying blood groups that makes use of convolutional neural networks (CNNs) and fingerprint-based analysis. There are distinctive ridge elements in fingerprint patterns that may be related to blood group traits, according to studies. The technology takes pictures of fingerprints, preprocesses them, and uses deep learning to identify distinguishing characteristics. CNN classifies 4 positive and 4 negative blood groups (A, B, AB, O) based on labeled fingerprint datasets. In addition, it offers an easier option than conventional blood testing. Dataset containing over 6000 fingerprint images have been used showing an overall accuracy of 88%. The model is rigorously trained and evaluated to guarantee excellent blood group prediction accuracy. This study emphasizes how AI-powered biometrics might be used in medical diagnostics, especially for emergency circumstances, healthcare, and blood donation administration. Furthermore, in emergency medical situations where quick blood group identification is essential, this strategy might facilitate blood transfusion procedures. Deep learning may be used for non-invasive medical diagnostics, as this study shows, opening the door for more AI-based advancements in biomedical research.