Large-scale database and data warehousing increasing, in the data collection process have the missing value are large problems. The problem of the missing value, it will seriously affect the quality of the analysis of data mining. So how to properly handle the issue of the missing value, will become an important topic. There has been the method that a lot of processing the missing value being put forward, among them with the concept of the association rule most is familiar. In this paper, the study take the association rule as the foundation, join the fuzzy set and taxonomy theories, and be used as to missing value completion as method according to the MVC technique. The study results show that minimum support and minimum confidence to the missing value completion of the correct value of a tremendous impact; In addition, to predict the fuzzy membership value of the change, the rate of correct data with the affected.