Due to the lack of a unified authentication model certain mistakes occurred in the use of the watermarking authentication methods. To clarify the confusion, authentication models of robust and fragile watermarking are developed respectively in the paper. Concrete algorithms are proposed to prove the models that different Discrete Wavelet Transform (DWT) domains are utilized to embed the watermarks and quantization method is presented with Just Notice Differences (JNDs) threshold as the quantization size. After the key technologies about the two methods are discussed, we detail the comparison of the two modes and recommend their respective application scenarios as applicable results.