With　the　rapid　expanding　of　Internet　of　Things,　context-aware　technology　has　become　a　research　hotspot.　At　present,　most　of　the　context　modeling　technologies　are　based　on　ontology,　however,　due　to　the　limitations　of　ontology,　the　ontology　framework　is　very　poor　for　the　uncertainty　reasoning.　A　few　experiments　have　been　introduced　fuzzy　theory　into　the　context-aware,　but　the　results　are　not　satisfying　and　many　problems　are　unsolved.　In　addition,　previous　studies　process　the　context　information　uniformly　without　considering　the　specific　characteristics　of　different　types　of　context　information.　In　this　paper,　we　have　designed　FuzWare　context-aware　middleware,　which　offers　a　context-aware　service　based　on　fuzzy　theory.　In　the　middleware,　the　context　information　is　divided　into　two　parts　according　to　the　information　sources.　The　result　shows　that　FuzWare　achieves　good　function　expansion　and　attains　accurate　reasoning.