<p>In recent decades, biometric technologies have become integral to security systems. However, the storage of original biometric data in databases presents a significant security vulnerability; if compromised, the biometric data is irretrievably lost. To mitigate this risk, research into cancellable biometric systems has emerged as a crucial area within biometric authentication. These systems transform biometric images into secure formats, rendering the recovery of the original data highly challenging. This paper presents a novel algorithm for cancellable biometric authentication, leveraging an optical chaotic system, Deoxyribonucleic acid (DNA) encoding standards, and specific operational processes. Our methodology involves the creation of cancellable templates for faces and fingerprints, utilizing chaotic maps and DNA encoding to encrypt biometric data prior to integration into the system. The algorithm employs a Piecewise Linear Chaotic Map (PWLCM) and a logistic map to generate initial values necessary for the encryption process, and DNA theory to encode the images. The process consists of several steps: first, generating a key image via the PWLCM; second, encoding the rows of both the plain and key images using DNA standards, with each row subject to a rule selected by the logistic map; third, applying DNA operations to combine the encoded images row by row, forming an intermediate image; and finally, processing this intermediate image for further encryption cycles. The robustness of our approach is validated through extensive testing on diverse face and fingerprint datasets. Comparative analysis with existing algorithms demonstrates that our approach achieves a notably low Equal Error Rate (EER) of 0.00047 and an Area under the Receiver Operating Characteristic (AROC) curve close to 1, highlighting its efficacy and security enhancements in biometric systems.</p>