With the development of Internet of Things, Vehicle Networking is increasingly getting widespread application, thus the higher requirements for accurate identification of the license plate; and the key of character recognition is the feature extracting and selection. The paper propose a method of feature extracting and selection based on wavelet moment and principal component analysis. At first, the character feature is extracted through wavelet moment. Then, selecting the feature through principal component analysis. Finally, eigenvectors are used as the input of BP neural network to complete character recognition. In the method, the global features and local features of image are catched, and has strong anti-jamming capability. The experiment shows that the method can achieve better recognition performance.