In　order　to　meet　the　diversified　visual　aesthetic　needs　of　the　public,　this　paper　constructs　a　digital　visual　model　based　on　art　history　theory.　Firstly,　the　SIFT　algorithm　is　used　to　extract　the　art　scene　feature　points,　and　the　convolution　operation　is　carried　out　by　Gaussian　function　to　form　a　three-dimensional　scale　space.　Then　the　system　reconstruction　technology　is　applied　to　realize　the　reconstruction　of　digital　images,　and　the　completeness　of　the　reconstructed　images　is　ensured　by　setting　visual　symbols.　Finally,　texture　segmentation　is　performed　by　combining　pixel　difference　features,　and　according　to　the　dynamic　distribution　of　the　corner　points　of　the　art　scene,　the　local　binary　fitting　method　is　used　for　the　information　enhancement　and　restoration　processing　of　the　art　image.　The　results　show　that　the　drawing　time　of　the　model　in　this　paper　is　always　kept　within　4s,　and　the　completeness　of　the　images　all　reach　more　than　99%,　and　the　highest　clarity　can　reach　0.99.　It　is　verified　that　under　the　guidance　of　the　art　history　theory,　the　attributes　of　the　digitized　visual　culture　have　been　innovated,　and　the　art　space　has　been　expanded.