This　paper　studies　how　smart　homes　provide　users　with　convenient　and　comfortable　services.　The　data　generated　by　users＇　daily　lives　can　accumulate　after　a　certain　period　of　time　and　can　form　a　huge　basic　data　set.　It　contains　many　valuable　information,　but　usually　the　information　values　of　these　data　sets　cannot　be　obtained　intuitively,　and　their　distribution　is　sparse.　How　to　use　them　effectively　and　mine　out　effective　information　to　provide　users　with　intelligent　life　services　are　always　the　goals　pursued　by　researchers.　In　this　paper,　machine　learning　technology　is　applied　to　smart　home,　and　a　smart　home　assisted　control　model　based　on　machine　learning　is　proposed.　The　model＇s　machine　learning　unit　will　fuse　the　data　of　the　two　aspects,　and　then　predict　the　status　of　home　equipment,　and　do　some　home　systems.　Behavior　feedback,　which　in　turn　helps　the　user　to　do　some　secondary　control　of　smart　home　system　operation.　Experiments　have　proved　that　the　assistance　system　designed　in　this　paper　has　a　high　accuracy　in　the　intelligent　control　of　equipment.　©　2020　IEEE.