The detection of counterfeit currency poses a significant challenge to financial security. This study presents a deep learning-based method for identifying fake currency using Convolutional Neural Networks (CNNs) based on a pre-trained VGG16 model. The dataset is composed of images of both real and counterfeit currency, which have been preprocessed and augmented to improve the model’s generalization capability. A transfer learning strategy is utilized, in which the VGG16 model, originally trained on ImageNet, is fine-tuned with additional dense layers for classification purposes. The model is trained using the Adam optimizer and binary cross-entropy loss function, achieving high accuracy in differentiating between counterfeit and genuine notes. Experimental results, including accuracy, loss curves, confusion matrix, and F1-score, demonstrate the model’s effectiveness in fake currency detection. The proposed system offers a reliable and automated solution for currency authentication, with potential applications in the banking and financial sectors.