Radar multi-target sorting is one of the important contents in electronic reconnaissance, and a precondition of implementing effective EW as well. The application of neural network to sorting radar multi-target is a new realm of electronic reconnaissance. The target parameters of time field, frequency field and space field are often used to sort radar multi-target. Kohonen neural network is capable of self-organizing and recognizing clustering center. A good result is obtained and the accurate value of clustering center is given out by applying one-dimension Kohonen neural network to radar signal sorting.