In the modern era of network popularization,various network protocols emerge one after another.At pres-ent,the most common is the IPv6 network protocol.Although it has optimized and solved some problems in the previous network protocols,there are still various security risks.In order to prevent and build a network defense system better,it is necessary to extract the characteristics of attack tree nodes in past attacks.Therefore,the node feature location of attack tree in IPv6 network attack event traceability is proposed.This method first uses the attack type to expand and obtain the at-tack tree model,and then calculates based on this model.Then it combines the deep learning network to extract the node features of the attack tree,and finally uses the DV-Hop method to locate the extracted node features of the attack tree.The proposed method not only has low energy consumption,but also has small error and good positioning performance,which can provide more effective data support for improving the network defense system.