Abstract: Radial noises of ships are very complicated. It is an imperative and difficult task to look for the new feature from the ship noises in the classification of underwater targets. Based on the technology of phase space reconstruction and the natural measure of chaos, a novel algorithm for extracting the feature from a ship noise is proposed in this paper. The results of the experiments indicate that the feature is satisfactory. It is practically demonstrated that the feature is very effective and it not only needs little time of the extracting calculation,but also has highly accurate rate of identification.