The rise of digital technology has made signature forgery a pressing concern across various domains, including financial institutions, legal documentation, and identity verification. This paper explores the performance comparison of various deep learning algorithms in terms of Signature Forgery Detection that leverages machine learning and image processing techniques to identify and prevent signature forgeries. The paper presents a comprehensive investigation into the performance of deep learning models, including CNN, YOLO, and VGG16, in the context of Signature Forgery Detection. These models were evaluated using the CEDAR dataset, a well-established benchmark for signature forgery detection. The study includes an in-depth analysis of training and validation results, visual representations of performance metrics, and confusion matrices for each model. In conclusion, the comparative analysis reveals that YOLOv8 emerges as the most effective model among the options, outperforming both CNN and VGG16.