This　paper　presents　Sniper,　a　knowledge-based　computer　field　search　engine　in　Semantic　Web.　Sniper　takes　WordNet　as　background　ontology　and　integrates　the　entities　in　the　semantic　documents　by　mapping　them　to　the　synsets　of　WordNet.　Sniper　returns　the　most　related　knowledge　in　computer　field　as　a　result　according　to　the　user＇s　query.　The　search　results　of　Sniper　are　displayed　in　the　form　of　a　list　of　entities,　including　concepts　and　instances,　which　are　extracted　from　heterogeneous　ontology,　so　the　result　is　close　to　the　user＇s　intention.　Compared　with　the　traditional　search　engines　and　the　current　semantic　search　engine,　Sniper　can　show　the　relevant　information　that　users　want　more　accurately.