This document proposes enhancing credit card transaction security in ATMs through a three-step verification process using password, fingerprint recognition, and one-time passwords (OTPs). It involves using KNN clustering to verify passwords, displaying a dummy balance if the password is incorrect, matching fingerprints to stored templates, and sending an OTP to the registered phone number if fingerprints match. This approach aims to strengthen user authentication and prevent fraudulent transactions using stolen cards and passwords guessed through brute force attacks. The system is analyzed in MATLAB and is found to provide improved security over existing one-step password verification methods.