Abstract: Birds are an important member of the ecosystem and the diversity of bird species has a self-evident effect on the ecological environment. Therefore, it is of great practical significance to protection of birds by identifying them through the sound of bird calls. In this paper, the bird sound signal is segmented by double-gate detection method of double parameters and used to find the specific frame of the sound starting point and ending point. A total of 5 features such as the short-term energy and short-term average amplitude in each segment of bird sound signal and the average value, contrast, and entropy in the short-time spectrogram are extracted, and then the support vector machine (SVM) with optimized parameters is used for birds species classification. The results show that compared with the ordinary support vector machine, the classification accuracy based on the support vector machine optimized by chaos cloud particle swarm optimization (CCPSO) is greatly improved. This method can effectively identify the birds and achieves the purposes of bird species protection, scientific research and ecosystem management.