This article focuses on the terrorist attacks, using social network analysis to study domestic terrorist activities. We have collected the text data of recent 30 years of domestic violence and terrorism in the World Terrorist Database legal district authoritative websites. The three elements like event, time and location are extracted for research. We used social network analysis software UCIET to build a multi-networks model of domestic terrorist activities, and adopt the meta-network decomposition method to decompose the multimode network into "time-location, time-incident, incident-location" 2-mode network, and converted it into a 1-mode network through rows and columns. Through the quantitative analysis of the centrality index of each subnetwork, and integrate of the two graphs and geographic information knowledge, we analyzed the characteristics of domestic terrorist activities from different perspectives. The result shows that the domestic terrorist activities clearly show the trend of spreading in large and medium-sized cities, which are vulnerable to the security of border areas, and the frequency of incidents on the timeline is wavy. These characteristics provide some practical significance for the estimate of terrorist activities and the evolution of the security situation.