The statistical analysis method based on probability distribution was used to further refine the test data of the turbine blade cooling effectiveness test (Reynolds number behind the gas grid, the total temperature ratio of gas to cool air, the flow ratio of cool air to gas and cooling effectiveness parameter, etc.). In order to obtain more accurate test data conclusions through parameter estimation and statistical inference. The normal probability graph, box plot and probability density distribution graph plus Gaussian fitting curves at the single state level of the above indirect measured values were plotted, and the degree of conformity of each measured data value to the normal distribution was quantitatively analyzed, with the outliers and confidence intervals pointed out. Based on the data processing results, the key factors that affect the probability distribution of the data were analyzed, which is of a strong reference value and practical value for this type of tests, and supplements and improves the blanks in the relevant fields.