Evaluation　and　Ranking　decision　making　units　with　independent　subsystems　are　important　contents　in　data　envelopment　analysis　(DEA)　research.　Applying　the　imaginative　worst　frontier　surface　concept,　this　paper　puts　forward　the　output　and　input　non-　radial　C(2)R　data　envelopment　analysis　model　which　has　two　stages　to　evaluate　and　rank　the　decision　making　units　with　independent　subsystems　based　on　the　non-　equal　proportion　expansion　or　reduction　thoughts.　The　method　in　this　paper　which　has　less　calculation　is　objective　and　optimal,　so　it　is　an　operable　and　practical　DEA　ranking　method.