With the continuous improvement and development of speech recognition technology, the numerous special purpose chips for food introduction speech recognition have been developed, thus, the practical products of speech recognition have been gradually appeared in the food market. This study will take the foundation of speech coding in Foreign-food Foreign-food speech system as the breakthrough point, by means of the interpretation of Speex, it discusses the design of the speech engine as well as the route of implementation so as to reduce tedious work in voice testing. Combined with the analysis of the basic structure of speech recognition in the Foreign-food system, it discusses the architecture of the speaker independent Foreign-food speech recognition learning system.