Reasonable　job　shop　scheduling　can　improve　the　production　efficiency　and　product　delivery　and　reduce　the　costs　and　energy　consumption.　The　quality　of　a　scheduling　scheme　mainly　depends　on　the　performance　of　the　used　algorithm.　Therefore,　several　researchers　have　attempted　to　improve　the　performance　of　algorithms　used　for　solving　the　flexible　job　shop　scheduling　problem　(FJSSP).　Currently,　the　genetic　algorithm　(GA)　is　one　of　the　most　widely　used　algorithms　for　solving　the　FJSSP.　However,　it　has　a　low　convergence　speed　and　accuracy.　To　overcome　these　limitations　of　the　GA,　a　novel　variable　neighbourhood　descent　hybrid　genetic　algorithm　(VNDhGA)　is　proposed　here.　In　this　algorithm,　a　barebones　particle　swarm　optimisation　(BBPSO)-based　mutation　operator,　a　hybrid　heuristic　initialisation　strategy,　and　VND　based　on　an　improved　multilevel　neighbourhood　structure　are　integrated　into　the　standard　GA　framework　to　improve　its　convergence　performance　and　solution　accuracy.　Furthermore,　a　real-number-based　chromosome　representation,　coding,　decoding,　and　crossover　method　is　proposed　for　maximising　the　advantages　of　BBPSO.　The　proposed　algorithm　was　tested　on　benchmark　cases,　and　the　results　were　compared　with　those　of　existing　algorithms.　The　proposed　algorithm　exhibited　superior　solution　accuracy　and　convergence　performance　than　those　of　existing　ones.