Abstract: Data analysis methods can analyze the potential connections or laws between data. In this paper, the method of multiple linear regression is used to select the measurement data of 27 blades on a certain leaf disc first. The data of 17 sets of parameters are measured by the three-coordinate measuring machine using the method of cross-section measurement, and the errors are calculated according to the theoretical values. And then, taking the error value of one parameter as the dependent variable and the error value of the remaining 16 parameters as the independent variable, the coefficients of the multiple linear regression equation were obtained through MATLAB software. Finally, according to the model test, the optimal model was selected, that is, TE THK the error value of the parameter is used as the dependent variable, and the other parameters are used as the independent variable. It is found that there is a strong linear relationship between the error value of the parameter TE THK and the error value of CV CONT MAX and CC CONT MAX, and a positive correlation; TE THK and LE THK (CA), WARP, SFT F/A, SFT CC/CV has the smallest linear relationship.