Bird species identification is a fundamental task in ornithology research, it sometimes can be a challenging and time-consuming process, particularly with large collections of audio signals. In this paper, we propose a mechanism based on audio signal processing, and machine learning to facilitate the identification of bird species. Our proposed mechanism involves two stages. In the first stage, we constructed an ideal dataset of sound recordings of different bird species which were subjected to various sound preprocessing techniques such as pre-emphasis, framing, silence removal, and reconstruction. In the second stage, the input is sent to our CNN model, and predictions are made.