In recent years, multimedia identification has become important as the volume of digital media has dramatically increased. With music files, one method of identification is audio fingerprinting. The underlying method for most algorithms is the Fourier transform. However, due to a lack of temporal resolution, these algorithms rely on the short-time Fourier transform. We propose an audio fingerprinting algorithm that uses a wavelet transform, which has good temporal resolution. In this thesis, we examine the basics of certain topics that are needed in understanding audio fingerprinting techniques. We also look at a brief history of work done in this field. We introduce a new algorithm, called the Hokua algorithm. We developed Hokua to take advantage of certain properties of the wavelet transform. The algorithm uses coefficient peaks of wavelet transforms to identify a sample query. The various algorithms are compared.