Abstract: The traditional methods of data processing in rupture strength experiment were the isothermal line method and Larson-Miller method. By isothermal line method, an equation should be built under each experiment temperature, so the method was tedious. In addition, Larson-Miller method should be discussed. Therefore, the multiple regression method had been developed. In multiple regression method, the multiple regression equation was constructed using the statistical relationship between temperature, fracture time and stress, and the rupture strength would be predicted by the equation. The errors to predict rupture strength were small and the precision of the method was high. The multiple regression equation could extrapolate and predict the time and temperature. It was not necessary to calculate the value and use of a logarithmic scale, and it was convenient and intuitive.