“Experts have proposed a bidirectional gate controlled cycle unit main steam temperature prediction model based on phase space reconstruction to address the problem of difficult prediction of main steam temperature fluctuations caused by peak shaving demand of thermal power units. The model is established through screening relevant variables, determining weight coefficients, optimizing hyperparameters, and other steps. It is compared and verified with models without relevant mechanisms. The results show that the model has the highest prediction accuracy and good dynamic characteristics, providing an effective solution to the problem of predicting main steam temperature fluctuations.”