This study compares three different pre-trained deep learning models specifically designed for fingerprint identification. The first model uses Convolutional Neural Network (CNN), the second includes Residual Network (ResNet), and the third employs the Visual Geometry Group (VGG) approach. The subsequent comparative assessment reveals the CNN-based model's superior performance, with an impressive F1 score of 96.5%. In contrast, the ResNet and VGG models achieve F1 scores of 94.3% and 92.11%, respectively. These findings highlight the CNN model's ability to accurately identify fingerprints. Furthermore, a comparative analysis is performed between the obtained results and those reported in recent studies using the same dataset. This analysis evaluates the performance of the proposed models and compares them to previous research, increasing confidence in the results. In conclusion, this study shows that in terms of fingerprint identification, the CNN-based model performs better than the other models.