With　the　continuous　popularization　of　smart　home　devices,　people　often　anticipate　using　different　smart　devices　through　natural　language　instructions　and　require　personalized　smart　home　services.　However,　existing　challenges　include　the　interoperability　of　smart　devices　and　a　comprehensive　understanding　of　the　user　environment.　This　study　proposes　Fraship,　a　framework　supporting　smart　home　service　personalization　for　end-users.　It　incorporates　a　runtime　knowledge　graph　acting　as　a　bridge　between　users＇　language　instructions　and　the　corresponding　operations　of　smart　devices.　The　runtime　knowledge　graph　is　used　to　reflect　contextual　information　in　a　specific　smart　home,　based　on　which　a　language-instruction　parser　is　proposed　to　allow　users　to　manage　smart　home　devices　and　services　in　natural　language.　We　evaluated　Fraship　on　a　real-world　smart　home.　Our　results　show　that　Fraship　can　effectively　manage　smart　home　devices　and　services　based　on　the　runtime　knowledge　graph,　and　it　recognizes　instructions　more　accurately　than　other　approaches.　©　2022　ACM.