Concerning the problem that the current network security situation assessment has characteristics of limited scope, single information source, high time and space complexity and big deviation in accuracy, a new network security situation assessment method based on Naive Bayes classifier was proposed. It fully considered multi-information sources and fusion of multi-level heterogeneous information, and had the features of rapidity and high efficiency. It dynamically demonstrated the whole security state of the network, which could precisely reflect the network security situation. Finally, the proposed method was verified using the reality data from network and its validity was proved.