Provided are an apparatus and method for transforming an application for a multi-modal interface. The apparatus includes: a parsing means for parsing an original application to generate a parsed structure of the original application; an application transforming/creating means for transforming the parsed structure generated by the parsing means; a multi-modal synthesizing means for synthesizing the contents created by the application transforming/creating means with multi-modal information of the user terminal; a multi-modal detecting means for detecting multi-modal information received from the user terminal and transforming the detected information into contents that can be used by the device; and an action mapping means for mapping the contents, which are received from the multi-modal detecting means and can be used in the device, to action information contained in the original application to generate an event of the device according to the action information.