The　strength　of　cement　is　one　of　the　important　indicators　of　its　Performance.　The　traditional　strength　detection　method　based　on　manual　timed　sampling　has　a　good　accuracy,　but　there　is　a　large　lag,　which　can　not　regulate　the　cement　production　process　in　time.　Machine　learning　en-ables　targeted　correlation　analysis　of　multi-dimensional　production　data　in　the　cement　industry,　for　example,　raw　material　and　product　fest　data,　microstructure　images　and　process　operating　Parameters.　The　lagging　problem　of　manual　testing　methods　can　be　solved　by　establishing　a　cement　strength　prediction　model　based　on　machine　learning.　This　paper　summarizes　the　cement　strength　prediction　model　based　on　machine　learning　by　sorting　out　the　basic　working　principles　and　advantages　of　commonly　used　algorithms,　and　discusses　its　application　effect　and　development　direc-tion,　with　a　view　to　providing　guidance　for　further　optimization　of　the　cement　strength　prediction　model　and　its　application　in　the　cement　industry.　©　2025　Cailiao　Daobaoshe/　Materials　Review.　All　rights　reserved.