This paper briefly introduces the design principles of the Spoken English Corpus of Public English Test System （SECOPETS） from the perspectives of the general scheme,the principle of data collection,the methods for data transcription and annotation as well as the development of the corpus processing tools.It also reports on the results of a preliminary study based on this corpus and has revealed,to some extent,the patterns of use and development of vocabulary and syntax in Chinese EFL learners＇ spoken English.