本研究提出多等級產品在相依製程下的監控方法。它能有效區分那個製程階段是失控的及診斷出引起所有製程平均值平移的非機遇因素和引起特定製程平均值平移的非機遇因素,以使失控的製程迅速回復穩定中。一個標準化的EWMA管制圖和各特定等級的EWMA管制圖被發展來管制標準化和不同等級的incoming資料。另外,一個標準化的選控圖和各特定等級的選控圖被發展來管制標準化和不同等級的out-going資料。這些管制圖的偵測力以模擬結果說明。最後,這些管制圖的經濟設計也以Markov chain和更新報酬方法推導出。 An approach to controlling multiple grades products produced on the multiple dependent process steps is developed. It not only distinguish which process step is out of control but also monitors the common assignable cause which affects overall process mean and the grade-specific assignable cause which affect grade-specific process mean effectively. A standardized EWMA control chart and grade-specific EWMA control charts are proposed to control grade-specific and standardized combined incoming data stream, respectively. Besides, a standardized cause-selecting control chart and grade-specific cause-selecting control charts are proposed to control grade-specific and standardized out-going combined data streams, respectively. A simulation results shows the effect of the proposed six control charts from statistical viewpoint. Finally, the economic scheme of the six proposed control charts is also proposed.