How to quickly and efficiently get real-time emergency data from a massive network is becoming a major problem in monitoring internet public opinion. In this paper, we propose a clustering algorithm based on simplified cluster hypothesis to improve system efficiency. Meanwhile, by analysis of the emergency’s characteristics under the dynamic web environment, we propose a multi-strategy quantitative analysis model for network emergency by researching on the following aspects: (1) Using field feature vector similarity to dividing emergency category. (2) Using Sum Limit method to calculate the emergency’s maximum effect range. (3) Using distribution to calculate the emergency’s timeliness. Finally, through analysis dynamic data from Tianya.cn, the experimental results prove the model is effective.