In order to solve the problem that the feature extraction of ship radiated noise in complex ocean environment is difficult, a method for feature extraction and classification of ship radiated noise based on variational mode decomposition, center frequency, complexity and support vector machine was presented. Four kinds of ship radiated noise signals were decomposed into several intrinsic mode functions with variational mode decomposition. In comparison, the center frequency and permutation entropy of intrinsic mode function with the maximum energy were taken as the characteristic parameters. The characteristic parameters acted as the input of support vector machine to distinguish the four kinds of ship. Results show that this method can realize the feature extraction of ship radiated noise, and it has higher recognition rate than the existing methods.