With development of the database, there are too many data on process variables and the manufacturing process for the
traditional statistical process control methods to identify the process variables related with assignable causes. Data mining
is useful in this situation and provides variety of approaches for improving the process. In this paper, we applied control
charts to monitor the process and if assignable causes are detected, then we applied the SVM technique and the sequence
pattern analysis to find out the process variables suspected. These techniques made possible to predict the behavior of
process variables. We illustrated our proposed methods with real manufacturing process data.