Abstract: This paper is basically for human voice recongnization for document recommendation and what problem we faced for document recommendation system. The system in known as automatic speech reorganization. The previous we used the written query for information gathering. The automatic speech recongnization machine is take input as human voice then it converts it to text. There are some methodologies used for this which will be discussed in this paper or document . The machine is actually worked as per voice given except that what voice it get from user. The system will give sentence as per human speak .The spoken words are converted into text query by use of ASR machine. The speech signals are given to system the training data .the automatic speech reorganization is worked on the acoustic model and pronunciation model.