In　this　paper　we　studied　the　technological　innovation　efficiency　of　the　Chinese　manufacturing　industry　for　each　province.　We　considered　R&D　employees　and　R&D　costs　as　input　variables,　and　new　product　sales　as　output　variable.　The　data　envelopment　analysis　(DEA)　approach　was　used　to　measure　efficiency　frontiers　that　identify　the　effective　or　ineffective　provinces　in　2005.　Constant　returns　to　scale　(CRS)　and　variable　returns　to　scale　(VRS)　DEA　models,　along　with　the　calculation　of　scale　efficiency,　are　employed　to　comparatively　analyze　the　efficiency.　Subsequently,　regression　analysis　is　applied　to　characterize　the　relationship　between　external　conditions　and　new　product　sales.　Finally,　a　discussion　of　what　we　found　and　suggestions　of　improving　the　technological　innovation　efficiency　are　addressed　in　the　conclusion.