An idea of establishing a more friendly and humanized application user interface is put forward, and an implement method of general purpose speech interface is presented.That makes Windows applications fulfil voice control conveniently. At the same time, several relatively advanced speech analysis methods are fully considered to improve speech recognition accuracy. What's more, the method of decision tree is used to select recognition algorithm dynamically according to voice signals that have different characteristics.That improves speech recognition accuracy greatly.In the end,a creative and ...