The project is targeted at the development of a model for the identification of bird species through their respective vocalizations. When datasets are involved with imbalanced class distributions, we prepare a sample set with sufficient audio data of those birds so that the model can be trained appropriately. Algorithms employed are Convolutional Neural Networks, Long Short-Term Memory networks, and WavNet in association with comprehensive feature extraction techniques. This study extracts three features from the audio: the zero-crossing rate, root mean square energy, and Mel-frequency cepstral coefficients, in that order. These are the most important features about species vocal identification. Data augmentation strategies used are noise addition and pitch shifting in order to enhance the performance of the model by diversifying the training samples. This enables achieving robust classification using a relatively small number of audio recordings per species. In total, our model is a success for correct predictions of bird species from audio input that can benefit biodiversity studies and ecological monitoring.