Abstract: Due to the non determinacy, the common nondestructive quantitative testing data could not meet the needs of the basic request of nondestructive evaluation and reliability safety analysis in engineering reality. In this paper, by the uncertainty evaluation method, the inverse regression analysis, the fuzzy theory and data fusion, the soft calibration method was built for uncertainty of measurement of nondestructive quantitative testing. The calculation and analysis results of some engineering example showed that based on the same accuracy of nondestructive testing instrument, the random and systematic uncertainty of measurements results of nondestructive quantitative testing data was clearly reduced by the soft calibration method, the quality of nondestructive quantitative testing data was raised clearly by overcoming some shortcoming of the common hard calibration. By the soft calibration method, the nondestructive quantitative testing data was capable of meeting the needs of the basic request for nondestructive evaluation and reliability safety analysis.