It is a very challenging task to develop effective process control methodologies for multi-operational manufacturing processes. Although Statistical Process Control (SPC) has been widely used as the primary method in the control of quality, it mainly serves as a change detection tool rather than a method to identify root causes of process changes. This paper proposes a systematic approach to overcome the limitations faced by SPC. In this method, a state space variation propagation model is derived from the product and process design information. The virtual machining concept is applied to isolate faults between operations, and further used in the root cause determination. The detailed methodology is presented, and a case study is conducted to illustrate and verify the developed diagnosis method.