Internet　of　Things　(IoT)　connects　various　kinds　of　sensors　and　smart　devices　using　the　internet　to　collect　data.　The　adoption　of　IoT　in　medical　care　field　will　bring　great　convenient　to　both　doctors　and　patients　for　effective　illness　monitoring　and　diagnosis.　Due　to　the　high　value　of　medical　data　and　the　openness　character　of　health　IoT,　the　protection　of　data　confidentiality　is　of　crucial　importance.　In　this　paper,　we　propose　a　novel　distributed　secure　data　management　with　keyword　search　system　for　health　IoT.　Since　the　patients　are　usually　managed　by　diverse　medical　institutions,　the　proposed　system　enables　distributed　access　control　of　protected　health　information　(PHI)　among　different　medical　domains.　On　the　other　hand,　the　accumulation　of　electronic　health　records　(EHR)　makes　effective　data　retrieval　a　challenge　task.　Our　scheme　could　provide　efficient　keyword　search　function　on　cross-domain　PHI.　For　the　resource　limited　devices　in　health　IoT,　it　is　an　essential　requirement　to　design　lightweight　algorithms　in　the　secure　data　management　system.　The　proposed　system　realizes　lightweight　data　encryption,　lightweight　keyword　trapdoor　generation　and　lightweight　data　recovery,　which　leaves　very　few　computations　to　user＇s　terminal.　The　security　of　this　system　is　reduced　to　the　decisional　bilinear　Diffie-Hellman　(DBDH)　assumption.　The　comparison　analysis　is　made　between　this　scheme　and　other　existing　systems.　The　extensive　experiments　on　both　laptop　and　smart　phone　platforms　show　that　the　proposed　scheme　has　greatly　improved　the　computation　efficiency　and　requires　much　less　communication　cost.