A　cross-shaped　array　which　was　composed　of　piezoelectric　sensors　was　used　for　the　excitation　and　reception　of　Lamb　wave　signals.　A　two-dimensional　multiple　signal　classification　(2D-MUSIC)　method　was　proposed　to　locate　and　detect　defects　in　aluminum　plates.　Firstly,　the　appropriate　excitation　frequency　was　selected　for　the　generation　of　a　single　mode　signal　to　avoid　the　influences　of　multi-mode　and　reduce　the　dispersion　of　Lamb　waves.　Then,　received　signals　were　segmented　by　2D-MUSIC　algorithm.　Finally,　reflected　signals　and　2D　steering　vectors　were　combined　to　locate　the　defects　in　the　aluminum　plate.　The　results　show　that　the　proposed　2D-MUSIC　algorithm　is　more　accurate　than　traditional　multiple　signal　classification　algorithm　for　defecting　locations　in　aluminum　plates.　©　2020,　China　Mechanical　Engineering　Magazine　Office.　All　right　reserved.