In this regard, this paper selects Poyang Lake given that the exploration object, adopts the transformer neural network system to ascertain the prediction product of each day scale of water amount of Poyang Lake, and explores its possible to forecast the water level change strategy of Poyang Lake. We Evaluate the simulation success of transformer m