Recognition of the radiated noises of ships is a very complicated and difficult job. In order to improve the classification performance and dependability of the single classifier, a kind of multi-classifiers decision making model and detailed algorithm based on many-person decision-makings theory was proposed. By adopting the eight categorized results of the BP neural network and the support vector machines which used the Welch spectrum, the linear predictive coding spectrum, the Burg spectrum and the perceptual linear predictive feature, the group decision-makings were done to recognize three different kinds of radiated noises of ships which were collected in many different operating conditions. Results show that the proposed fusion model and algorithm are feasible and the statistical right recognition probability arrives at 96.44% and its classification performance is superior to the performance of the single classifier. The method can be applied to the underwater acoustic target recognition system.