In the past 40 years, scientists who studied on whales had to manually match whale individuals with the photos they had gotten in order to monitor the size, health of humpback whale groups. However, this was relatively inefficient for the process of building up a thoroughly global database. Therefore, it is inevitable to develop an advanced technique to identify whale individuals with their tale photos. The effective solution to this issue will be training convolutional neural networks with multiple layers to let models categorize photos. We actually not only trained neural networks but also compared different models that were applicable to identify images. With a limited amount of training data, we were able to identify some whales with the test whale tail images.