Speech is one of the natural forms of communication. In the real world, speech signals are nonlinear, non-stationary in nature. Speaker identification is the task of determining who is talking from a set of known voices or speakers. The performance of speaker identification systems still need of improvement in term of text-independent speaker identification and suitable modeling techniques for voice feature vectors. The Mel frequency cepstral coefficients technique is used to analyze the speech signal and extract the features of speech signal and also Vector Quantization technique is a lossy data compression technique which is used to generate the codebooks which are modeled during training phase and compared with the codebooks of the reference samples stored in the speaker database during speaker identification process. In vector quantization technique by increase in codebook size increases the performance of speaker identification process.