Customer value is the basis of customer relationship management. This paper applies Na?ve Bayes model to analysis data mining of clients' transaction data and the basic information. It uses Na?ve Bayes model of Weka to train, test and check customer transaction data, which is to construct customer value analysis model, then evaluate itself, finally establish a highly efficient data mining model. The model classifies customer and value analysis so that it can identify different characteristics, different value of the customer base, all that can help provide personalized service according to different levels of customer value.