The variable stator vane (VSV) is an important mechanism for the airflow control of the aviation engine compressor, which can expand the stable working range of the compressor and improve its pneumatic performance.Based on the massive LEAP engine data stored in the aircraft quick access recorder (QAR), this paper established a method to predict the VSV Angle of LEAP engine based on 1DCNN-Resnet-LSTM model. The evaluation results of the model show that the error of the simulated output is small compared with the actual output, and the change of the simulation trend can fit the actual data well, indicating that under the background of today's big data, it is feasible and significant advantages to use the deep learning algorithm to predict the VSV Angle and explore its control law.