Biometric authentication is critical for secure access in diverse domains, emphasizing the need for high accuracy to prevent security breaches. Multimodal biometrics fuse traits to fortify security, vital across diverse domains, surpassing unimodal systems’ vulnerability. This paper introduces a multi-modal biometric recognition using palm print and iris datasets. In our approach, Gabor filters are used for feature extraction with an innovative fuzzy bit transition coding (BTC). Followed by, a novel hybrid Jellyfish Search Optimization (JSO) and Cat Swarm Optimization (CSO) algorithm for feature selection of iris and palm print modalities. Firstly, the palm print images undergo feature extraction using the fuzzified BTC and the preprocessing of iris images is done using a Convolutional Neural Network(CNN) to refine the dataset and identify regions of interest (ROI). After that, extracted features are then selected using our hybrid JSO-CSO algorithm, which are then utilized for classification in a deep learning based ResNet model. The two biometric modalities process in parallel, and their classification results are fused using weighted score-level fusion for the final recognition result. This method improves feature interpretability as well as classification performance; it has been experimentally validated against MMU iris database, CASIA-palm print dataset and a custom multimodal database resulting in peak accuracy of 97.6%. The proposed hybrid JSO-CSO algorithm outperforms other techniques in terms of precision and convergence, thus providing a robust solution for multi-modal biometric recognition.