An innovated feature extraction algorithm by using discrete wavelet transform (DWT) for audio fingerprinting was proposed in this study. In this study, we attempt to reduce minimal length of audio needed for identification (i.e. granularity) for increasing flexibility in dealing with various requirements of an audio fingerprinting system. Moreover, both the robustness and reliability issues of fingerprinting system are addressed. The experimental results show that the proposed method is not only reliable but also robust against various signal degradations. Furthermore, the granularity of fingerprint is smaller than previous work.