This　paper　analyses　the　characteristics　of　flexible　job　shop　scheduling　problem　(FJSP),　and　proposes　a　genetic　simulated　annealing　algorithm　to　solve　the　scheduling　problem.　Considering　the　balance　of　the　utilization　rate　of　each　machine,　the　maximum　utilization　rate　of　machine,　the　maximum　completion　time　and　other　performance　indicators　are　more　reasonable,　a　genetic　simulated　annealing　algorithm　combining　global　search　and　local　search　is　designed.　Through　the　design　of　reasonable　crossover　operator,　the　global　search　ability　of　the　genetic　simulated　annealing　algorithm　is　improved.　In　addition,　the　precocity　is　improved　by　the　simulated　annealing　algorithm,　and　the　quality　of　understanding　is　improved.　And　finally,　the　effectiveness　and　feasibility　of　the　algorithm　is　verified　by　simulation　experiments.　　©　2020　Author(s).