As　the　search　technology　rapidly　developed,　nowadays,　main　search　engines　are　already　able　to　meet　users　basic　search　desire.　However,　current　search　algorithms　or　methodologies　mostly　depend　on　keywords　matching　process,　which　could　be　effective　for　text　search　while　not　efficient　for　keywords-lacking　or　non-text　search　scenarios.　This　paper　summarizes　the　solution　adopted　by　current　search　engine　vendors,　and　introduces　a　new　approach　that　attaches　content　description　index　based　on　RDF　standard　to　web　images　in　order　to　achieve　converting　unstructured　information　search　into　structured　information　search.　By　injecting　the　Web　2.0　feature-Wisdom　of　Crowds,　the　engine　has　the　characteristics　of　self-learning:　as　users　amount　increases,　the　knowledge　base　for　semantic　content　of　images　accumulates,　which　makes　the　search　engine　more　and　more　intelligent　for　search　and　semantics　reasoning.