This　paper　proposes　an　optimization　method　based　on　the　Black　Hole　Cuckoo　Search　Algorithm　(BH-CS)　to　improve　the　accuracy　of　the　weighted　similarity　measure　in　the　case-based　reasoning　(CBR)　model.　First,　take　the　root　mean　square　error　of　the　case-based　reasoning　prediction　model　as　the　fitness　function.　Secondly,　use　the　Levy　flight　of　the　Cuckoo　Search　algorithm　to　update　the　feature　weights　in　the　weighted　similarity　measure　method　and　evaluate　the　optimal　weights　from　them.　Then,　randomly　generate　new　feature　weights　with　some　probability.　Finally,　use　the　Black　Hole　algorithm　to　optimize　feature　weights　further　to　obtain　optimal　weights　and　optimize　the　case　similarity　measure.　The　optimization　method　was　tested　using　UCI　standard　data　set.　The　results　show　that　the　BH-CS　algorithm　has　an　advantage　over　other　algorithms　in　improving　the　accuracy　of　case　similarity　measures　and　can　effectively　improve　the　prediction　accuracy　of　the　CBR　model.