The　multiple　objective　functions　or　constraints　of　dynamic　multi-objective　optimization　problems　(DMOPs)　generally　vary　over　time.　The　multi-objective　evolutionary　algorithm　based　on　decomposition　(MOEA/D)　not　only　tracks　the　entire　Pareto　optimal　front　(POF)　but　also　has　good　performance.　In　effect,　a　decision　maker　(DM)　is　only　cared　about　part　of　POF　in　different　environments,　which　is　called　region　of　interest　(ROI).　To　solve　this　problem,　preference　weight　vector　adjustment　strategy　based　dynamic　MOEA/D　(MOEA/D-DPWA)　is　proposed.　Firstly,　the　reference　vector　adjustment　strategy　is　put　forward　to　bring　in　the　preference　information　of　DM　into　DMO.　Then,　the　dynamic　response　strategy　is　implemented,　which　is　designed　to　regenerate　the　population　near　to　ROI　when　environmental　dynamics　change.　Finally,　experimental　results　indicate　that　MOEA/D-DPWA　is　valid　in　tackling　DMOPs.　©　2023,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.