In　order　to　propose　an　effective　machine　learning　prediction　model.　This　paper　constructs　an　ensemble　learning　prediction　model　based　on　a　support　vector　machine　(SVM),　classification　regression　tree　(CART),　and　multi-layer　perceptual　base　(MLP),　applying　R^2　and　RMSE　to　measure　the　accuracy　of　the　prediction　model　and　compares　three　The　performance　of　the　base　learner.　The　results　show　that:　(1)　the　RMSE　and　other　parameters　of　the　ensemble　learning　are　small,　the　R2　is　larger,　and　the　prediction　effect　is　significantly　higher　than　that　of　the　basic　learner　(2)　the　ensemble　learning　model　based　on　SVM-CART-MLP　has　three-time　intervals　R2　and　RMSE.　Performance　0.98　vs　0.135.These　models　have　strong　general　and　scalable　conditions.　©　2023　IEEE.