Automation　in　robotic　endoscopy　surgery　is　an　important　but　challenging　research　topic.　The　flexible　deformation　of　the　endoscope　makes　it　difficult　for　surgeons　to　steer　it,　especially　in　a　narrow　workspace.　The　safety　of　endoscope　steering　has　been　a　major　concern　in　a　robotic　endoscope.　This　study　presents　a　safety-aware　control　framework　to　enable　a　robotic　endotracheal　intubation　system　(RNIS)　to　automatically　steer　the　flexible　endoscope　during　nasotracheal　intubation　(NTI).　With　the　endoscopic　image　as　feedback,　we　used　the　fusion　of　image　intensity　and　optical　flow　measurements　to　detect　the　lumen　center.　A　velocity-based　orientation　controller　is　designed　to　adjust　the　orientation　of　the　endoscope　tip　to　track　the　lumen　center.　To　ensure　surgical　safety,　the　relative　depth　from　the　endoscope　tip　to　the　lumen　surface　is　calculated　from　endoscopic　images.　With　the　relative　depth　as　feedback,　the　feed　motion　of　the　endoscope　is　controlled　to　prevent　the　collision　between　the　endoscope　tip　and　its　surrounding　tissue.　The　position　and　orientation　of　the　endoscope　tip　can　be　adjusted　by　the　RNIS　during　NTI.　Finally,　extensive　experiments　based　on　a　training　manikin　were　performed.　The　results　indicate　that　the　proposed　control　framework　enables　the　RNIS　to　successfully　steer　the　flexible　endoscope　along　the　upper　respiratory　tract　with　high　safety.　The　performance　of　lumen　center　detection　is　good　according　to　the　annotation　of　experienced　users.