A technique of identical a set of speech utterances for identifying and verifying speakers using singular value decomposition (SVD) is accessible. The set of stored reference parameters yielding the largest ratio of singular values when mapped with the parameters of the utterance of an unknown speaker is considered closest to the unknown. This research examines the feasibility of integrating a speaker verification model in a voice authentication system. The speaker verification model is used to verify the identity of the speaker. Speaker voice to be extracted using familiar MFCC feature Extraction, after completing extraction process decomposition is exposed using SVD and also mapped with the parameters of the utterance of an unknown speaker. From the experiments, validation accuracy of the speaker verification model will be tested.