Abstract: In order to study the difficultly measurable changes of the internal energy, the fluid operating state and the reaction composition in the steam generator, which are caused by the damage of the heat transfer pipeline in the steam generator of the sodium-cooled fast reactor, the time series model forecast analysis method was introduced. The noise signal generated by the steam generator experimental bench under simulated leakage conditions was analyzed by model and forecasted based on the time series model. The specific time of heat transfer pipeline leakage was determined by comparing the forecasted signal with the original signal. The results show that the noise signal generated by the sodium-water reaction can be modeled and analyzed effectively by the time series model forecast analysis method, the specific time can be identified quickly and accurately when the leak occurs according to the forecast signal.