The invention provides a fingerprint feature extraction method based on wavelet transformation. The method comprises the following steps that a fingerprint image is extracted and binarized; discrete wavelet transformation is carried out on the binarized fingerprint image to extract features from the fingerprint image; detailed fingerprint features are converted into characteristic vectors of the wavelet domain; and the obtained characteristic vectors serve as final characteristic vectors, and are identified or stored by being compared with a fingerprint library. The extraction method of the invention can be used to avoid tedious calculation of preprocessing effectively, reduce the feature obtaining time as well as time in comparison with the fingerprint feature library, accelerate identification and improve the instantaneity of a fingerprint identification system.