Given　the　increasing　social　needs　for　high-quality　health　and　medical　services　at　a　low　cost,　smart　health　has　gained　significant　attention　as　the　leader　in　achieving　national　happiness　and　next-generation　growth　engine　based　on　ICT　convergence　technology.　To　meet　the　needs　of　home　and　primary　healthcare,　this　paper　proposes　an　application　scheme　based　on　machine　learning　and　similarity　calculation　algorithm　for　home　and　primary　healthcare.　Users　can　move　freely　at　home　at　any　time　and　obtain　accurate　human　physiological　parameters,　good　medical　services,　and　personalized　doctor　recommendations.　The　scheme　can　be　used　for　home　and　primary　healthcare,　and　has　good　practical　value.