Personalized　recommendation　is　an　important　technical　means　to　deal　with　information　overload　and　meets　users＇　personalized　customization　needs.　In　response　to　the　insufficient　extraction　of　user　historical　features　and　cold　start　problem　in　traditional　movie　recommendation　systems,　we　introduce　an　effective　personalized　movie　recommendation　model　named　Movie　join　with　Starring　and　Ratings　(MSR),　based　on　attention　and　gate　mechanisms.　The　paper　uses　multi-head　attention　mechanism　to　capture　the　intrinsic　relationship　between　various　data　fields　in　user　viewing　records　and　obtains　user　features　through　the　Basic　Information-Rating　Joint　Attention　Network　(BRJA);　the　gate　mechanism　is　used　to　integrate　the　basic　information　and　reputation　rating　of　the　movie　into　the　movie　representation　vector,　obtaining　candidate　movie　features.　The　model　can　provide　good　recommendations　even　with　limited　user　information,　effectively　addressing　the　cold　start　problem.　Comparative　experiments　on　the　public　dataset　Movie　Lens　and　ablation　experiments　on　key　modules　have　demonstrated　the　effectiveness　of　the　MSR　model.　©　2023　IEEE.