Abstract: To solve the problem of the quality management system of the discrete manufacturing enterprise is segmented,decentralized and lacking of large-scale quality data analysis under the traditional concept, this paper analyzes the quality data from the perspective of the whole life cycle,it have discussed the data acquisition, data preprocessing and data conversion before data mining,and have constructed the ETL data preprocessing model. Then, studied the concrete data mining application method, put forward the overall data mining architecture of product life cycle quality data,and analyzed the data layer, method layer, knowledge layer and application layer contained in the architecture.