Abstract: The quality of a diagnosis system impacts the living quality and even life of patients. But regular software testing tools have difficulty in testing and evaluating the quality of a complex diagnosis system. Modeling and evaluation of a diagnosis system based on the Hierarchical Colored Petri Net before its development is highly necessary, feasible and forward-looking and can help detect loopholes and deficiency in the system and can provide informative support to the diagnosis of illness and designers of medical information system at large. Different treatment options are generated for patients with similar physical conditions but varied willingness in emulation tests on CPN tools 3.4.0. Upon comparison with 20 actual cases of the Henan Provincial People's Hospital, such system generated treatment options are substantially in line with that of the actual cases. A Gastric Cancer Dynamic Auxiliary Diagnosis System is designed in accordance such modeling and algorithm as detailed in this thesis and is proven to function regularly.