Abstract: To eliminate a wrong recognition of speech and reduce processing load to be applied to a computer in a simulation apparatus. The simulation apparatus includes a speech recognition section for referring to dictionary data in which a plurality of phrase data are recorded to specify phrase data, a dictionary selection section for selecting the dictionary data to be referred to by the speech recognition section in response to a mode, and a monitor and a motor for being operated based on an operation of rider, speech and the phrase data. The speech recognition section confirms whether or not sound inputting is valid. If the sound inputting is valid, then a speech recognition mark is displayed on a screen. One of the dictionary data is selected based on the mode. Voice of the rider and the dictionary data are referred to for searching for phrase data conforming to the recognized speech.