Aimed at the difficulty in equipment fault prediction, a new model has been established based on the Grey GM(1,1) model and LRM combined with principles of validity. It is used to estimate when the system reaches the upper limit of the failure data by simulating and predicting the original data of the equivalent interval measurement, and to infer the system failure time according to the results. At the same time, the metabolic method is introduced to improve the prediction accuracy. Finally, the output voltage of a certain type of radar transmitter data is taken as an example to verify the effectiveness and practicality of the model in failure prediction.