Aiming at the problem that aircraft status data cannot be monitored in real time due to transmission abnormalities during transmission to the ground, a flight status prediction method based on random forest (RF), temporal convolutional network (TCN), sparrow search algorithm (SSA), and Informer model was proposed. search algorithm (SSA), and Informer model for flight state prediction. Random forest algorithm was used to analyze the feature importance of real flight data, and the filtered parameters were used as input variables for prediction. TCN was introduced to make up for the problem of information loss or blurring of information in the Informer model when dealing with complex flight data sequences, and TCN was able to capture the local time features and short-term dependence effectively through its convolutional structure, and fusion of TCN and Informer's feature matrix was carried out to capture the flight state prediction simultaneously. The fusion of TCN and Informer's feature matrix can capture both local and global information in the flight state sequence, thus improving the accuracy and reliability of prediction, and the sparrow search algorithm was used to optimize the hyperparameters of the model in order to obtain better prediction performance. Experimental validation was carried out using the flight pitch angle, roll angle, and yaw rate as the prediction objects, and through comparison and analysis with other models. The experimental results show that the proposed model has good prediction performance in both single-step prediction and multi-step prediction tasks, and it can provide a reference for the prediction of aircraft state.