Sharks are cartilaginous fish that are widely hunted because they have high economic value. Overfishing and trade have resulted in this species being threatened with extinction and have been included in several categories of the IUCN Red List. Information on the types of sharks that landed in Sungailiat Bangka PPN is still very limited due to the difficulty of morphological identification, so it is necessary to identify them using molecular methods. Therefore, the researcher produced an image recognition program in sharks using the Convolutional Neural Network algorithm, which is a convolutional activity by combining several layers of preparation, by utilizing several components that move together and are motivated by a biological sensory system. The shark images used are basking, blacktip, blue, bull, hammerhead, lemon, mako, nurse, sand tiger, and thresher. The implementation of shark image recognition is carried out using 2 test models, namely the Sequential model and the VGG16 on top model running on the Google Collaboratory application, and Keras. The test data in this study were 1089 training data images and 1073 test data images which resulted in an evaluation value with an accuracy value of 86.58% and a loss value of 0.701 on the Sequential model and an accuracy value of 91.80% and a loss value of 0.0355 on the on top model. VGG16.