The study explores bird species classification by utilizing Deep Convolutional Neural Networks (DCNN) based on the GoogLeNet framework. It highlights the challenges of variations in bird size, form, and color. The method involves converting bird images into grayscale to create autographs, which are then analyzed to predict species. The research employs the Caltech-UCSD Birds 200 (CUB-200-2011) dataset, containing 500 labeled images for training and 200 unlabeled images for testing. The DCNN model achieved a prediction accuracy of 88.33%, leveraging GPU technology on a Linux system using TensorFlow and an NVIDIA Geforce GTX 680 graphics card.