Abstract: This study uses S-H-ESD algorithms to monitor the outliers in the process of dissemination of public opinions on wild animals, which can more accurately find the segmentation points of each stage, and provide a certain empirical basis for exploring the changing characteristics and internal laws of each stage in the process of disseminating online public opinions on wild animal. It is helpful to grasp the correct guidance of public opinions for wildlife protection. This study uses S-H-ESD algorithm to divide the stages of wildlife-related network events with mass public opinions. The results show that the life cycle of communication of wildlife-related network public opinions can be divided into germination period, outbreak period, mitigation period and decline period. On this basis, according to the characteristics of each stage, a complete set of strategies and suggestions for "monitoring-intervention-guidance-evaluation" are proposed to deal with the crisis of network public opinions on the wildlife. Finally, we propose the following suggestions: ①improve the online public opinion risk assessment system concerning wild animals, and classify the risk levels of different types of wild animals, strengthen the risk management and control of public opinions on the wildlife network; ②establish an early warning and supervision system for wildlife-related network public opinions to scientifically prevent the problem of "conflict between man and wildlife".