Beginning with the analysis of the problems that the chinses proficiency testing has brought about in recent years. the purpose of this paper is explore the rather reliability and test validity of FLEX developed by Hankuk University of foreign studies. This study investigates the potential roles in the Generalizability theory in the validation of a performance-based chinses test. Data for this study come from Hankuk University of foreign studies`s FLEX. There is no doubt that the FLEX reform will change the traditional chinese way of teaching, teachers` dominating the classes by explaining the language points. Therefore, teachers should pay more attention to the formative test, such as diagnostic tests as the various learning periods, so that the teaching plans and methods can be adjusted in time.