The pervasive issue of counterfeit currency poses a significant threat to global economies, leading to substantial financial losses and eroding trust in financial institutions. While traditional methods of counterfeit detection rely on visual inspections and security feature verification, the advent of sophisticated counterfeiting techniques necessitates more advanced solutions. This study explores the application of machine learning (ML) algorithms to the problem of currency authentication, utilizing a dataset comprising images of banknotes. We employed a convolutional neural network (CNN) architecture, tailored for feature extraction from image data, which has demonstrated exceptional performance in pattern recognition tasks. Our proposed model introduces a novel preprocessing algorithm that enhances the distinctive characteristics of genuine and counterfeit notes, improving CNN's classification efficacy. We benchmarked the performance of our model against traditional ML classifiers, such as Support Vector Machines (SVM) and Random Forests. The analysis reveals the strengths of our model, including its robustness to variations in image quality and its ability to generalize from limited training data. However, limitations were observed in terms of computational intensity and the need for a substantial dataset to optimize model parameters. The findings underscore the potential of automated, ML-driven techniques in revolutionizing the detection of counterfeit currencies, presenting a scalable and effective solution to a complex challenge.