Abstract: The measurement uncertainty of the analytical results of sulfides in groundwater samples by para-amino aniline photometric method was evaluated using continuous propagation model of uncertainty. At the key step of standard curve fitting, the authors adopted the double-error curve fitting in comparison with the least square method. The results show that, for the higher concentration samples, the results of double-error curve fitting are more accurate, and when double-error curve fitting is adopted to quantify the actual samples, the appropriate concentration range of standard curve should be selected, in which the concentration of the samples falls into the range of higher concentration, so that the results are more accurate and reasonable.