Cardiovascular　disease　is　the　number　one　killer　threatening　human　life　and　health.　Globally,　about　40%　of　deaths　are　caused　by　cardiovascular　disease.　This　paper　builds　a　model　based　on　logistic　regression　method　and　similarity　evaluation　based　on　distance　formula　to　predict　cardiovascular　disease.　The　data　in　this　article　comes　from　the　kaggle　platform.　In　the　experiment,　we　used　the　logistic　regression　algorithm　to　build　the　first-stage　model.　The　accuracy　and　recall　rate　of　the　model　reached　88.5%　and　89.2%.　Then　using　the　similarity　evaluation　method　based　on　Euclidean　distance,　we　optimize　the　data　that　cannot　be　processed　in　the　first　stage.　Finally,　a　complete　classification　model　is　obtained.　Finally,　we　bring　the　test　data　into　the　complete　model.　The　recall　rate　of　the　model　finally　reached　98%,　and　the　prediction　accuracy　rate　also　reached　98%.　©　2021　IEEE.