Aero-engine assembly is the key part of the manufacturing process of aero-engine which are composed of many complex productions. And disturbances appear frequently in the manufacturing process, such as assembly time fluctuation, unqualified rework, etc. Aiming at the scheduling problem of the aero-engine assembly line under uncertain environment, the adaptive scheduling method based on a gate recurrent unit(GRU) neural network is studied. The scheduling method includes two core modules: Disturbance identification and scheduling scheme adjustment. In the disturbance recognition module, a GRU neural network is built to identify progressive disturbances in rolling time windows. Once disturbances are recognized, the scheduling adjustment module will be driven, and the best scheduling rules on current scenarios based on the GRU neural network are output, which are used to update scheduling scheme. Finally, a case study is carried out for an aero-engine assembly line. The comparative experiments show that the proposed method can effectively improve the equipment utilization rate and daily productivity of the assembly line.