Based on prediction method, a novel method of automatic interpretation for problem of the slow-varying telemetry parameters changing suddenly is proposed. Firstly, extract the envelope of the flight data and then filter the envelop. Secondly forecast the filtered envelop using linear prediction method and prediction method based on GM (1,1) model. Finally, according to the estimated value and standard deviation of the parameters, the statistical analysis of the target data is carried out, where the parameters are determined to be abnormal or not according to the analysis result, so as to realize the automatic interpretation function of the parameters. The test results show that the prediction method based on GM (1,1) model can effectively identify the mutation or pulsed parameters in the slow-varying telemetry parameters Compared with the traditional method of the artificial interpretation, this method can effectively improve the efficiency and accuracy of the telemetry parameter interpretation.