We analyzed the effects to the mechanism of speech utterance and production from auditory system by fitting a stochastic time series model to pitch and power data extracting from recorded speech time series. It includes that 1. The noise contributing to each fluctuation of pitch and power influences each other, 2. Transformed Auditory Feedback (TAF) method might make a little bit changing the normal speech condition. In this analysis, we can support the results obtained from TAF method. Furthermore, we can investigate mutual relationship between speech parameters which the TAF method could not find.