On a basis of the DA/AD conversion analysis of experimental environment, the factors that the difference occurs in the audio signal in a relatively severe DA/AD conversion environment are presented. The most important three reasons that contribute adversely to the audio watermarking—noise interference, time-domain drift and amplitude difference, are analyzed and summarized. The quantization-based audio watermarking algorithm is used, in which the appropriate frequency-domain is selected and BCH error correcting codes for the watermarking information is achieved. The result shows that, compared with the existing algorithms, the method has a better capability for resisting against the DA/AD conversion, as it is applicable to various types of music.