This research introduces a hybrid prediction model that combines convolutional neural networks (CNN) and long short-term memory networks (LSTM) to enhance the precision of seismic response analysis for high-speed railroads. The model uses quasi-distributed fiber optic gratings to monitor the track plate, rail, base plate, and beam during seismic activities. By strategically placing seven grating monitoring points on each fiber, the model predicts the central point's response using data from peripheral gratings. The CNN extracts initial features from a continuous feature map formed via a time-sliding window, which are then sequenced for the LSTM network for prediction. Empirical results show that the model has an RMSE of 0.3753, MAE of 0.2968, and R² of 0.9371, demonstrating its effectiveness in earthquake response analysis for rail infrastructures. The study also discusses the limitations of the model, particularly its computational intensity in real-life scenarios, and highlights the need for further exploration to ensure real-time applicability.This research introduces a hybrid prediction model that combines convolutional neural networks (CNN) and long short-term memory networks (LSTM) to enhance the precision of seismic response analysis for high-speed railroads. The model uses quasi-distributed fiber optic gratings to monitor the track plate, rail, base plate, and beam during seismic activities. By strategically placing seven grating monitoring points on each fiber, the model predicts the central point's response using data from peripheral gratings. The CNN extracts initial features from a continuous feature map formed via a time-sliding window, which are then sequenced for the LSTM network for prediction. Empirical results show that the model has an RMSE of 0.3753, MAE of 0.2968, and R² of 0.9371, demonstrating its effectiveness in earthquake response analysis for rail infrastructures. The study also discusses the limitations of the model, particularly its computational intensity in real-life scenarios, and highlights the need for further exploration to ensure real-time applicability.