An　intersection　control　approach　based　on　case-based　reasoning　(CBR)　is　put　forward　in　this　paper.　CBR　solves　new　problem　by　reusing　solution　of　similar　old　problem.　The　strength　of　CBR　lies　in　its　ability　to　simulate　the　decision-making　process　of　a　human.　Case　structure　is　presented　and　decision　tree　is　used　to　organize　the　cases.　The　most　similar　case　is　obtained　using　the　nearest　neighbor　algorithm.　A　simulation　study　is　conducted　on　an　intersection　and　numerical　results　demonstrate　the　effectiveness　of　the　proposed　approach.