: aiming at the problem of low detection efficiency and longer detection time in the current DoS attack detection algorithm, it proposed a DoS attack detection algorithm based on S-Kohonen. With this algorithm could realize and quantify the segmentation of network traffic data packets, and extracted the cumulant feature effectively, so as to use cumulants in the process of DoS attack detection. Through the comprehensive analysis of the modern intrusion detection data sets, it would implementfull detection of DoS attacks. Compared with the traditional anomaly detection algorithm based on the entropy of network traffic, this algorithm could effectively improve the detection accuracy and shorten the detection time.