Technological　innovation　is　an　increasingly　important　determinant　of　the　Chinese　manufacturing　industry;　while　improving　R&D　efficiency　plays　a　significant　role　in　the　field　of　technological　innovation.　In　this　paper,　the　SFA　(Stochastic　Frontier　Analysis)　as　well　as　the　CRS　(Constant　Return　to　Scale)　and　VRS　(Variable　Return　to　Scale)　methods　of　the　DEA　(Data　Envelopment　Analysis)　models　are　used　to　evaluate　the　relative　efficiency　of　a　sample　from　the　manufacturing　industry　in　China.　Given　the　theoretical　differences　between　the　two　techniques,　this　paper　gives　an　empirical　comparison　of　the　two　models　using　the　calculation　of　efficiency　scores.　The　frequency　distributions　of　the　measurements　as　well　as　Spearman　correlation　coefficients　are　presented　to　examine　the　agreement　of　these　two　approaches;　Some　discussions　to　R&D　efficiency　results　are　addressed　in　the　conclusion.