To eliminate the need for mouse operation and to improve an effect of a speech recognition device by automatically display words stored in a word storage means for recognition for helping a speech input to the speech recognition device by user.

A word storage means 11 for recognition stores all words as recognition object of an input speech. Based on the feature of the input speech, a speech recognition means 12 determines a most similar word to the input speech from the words for recognition, and outputs it as a recognition result. Further, a display means 13 automatically displays the words stored in the word storage means 11 for recognition to help the speech input to the speech input system by the user. For example, when a word is found which should be inputted in the words displayed on a display screen, the word is speech- inputted. The input speech is collated with all the recognition words stored in the word storage means 11 for recognition based on the feature of the speech, a most similar word to the input speech is outputted from the speech recognition means 12 as a result of the recognition.