This paper compares two techniques to assess uncertainty on emission measurements. The first one, described in ISO 14956, gives an appropriate procedure to establish uncertainty budgets from systematic assessment of factors influencing the result. The second approach consists in the quantification of the fidelity of the method during inter and intra-laboratory field experiments set out according to ISO 5725-2. The comparison has been carried out for two reference methods described in the drafts prepared by CEN/TC264/WG16. For the SO2 manual reference method Uncertainty budgets lead to very realistic overall uncertainty comparable to the repeatability and reproducibility confidence intervals determined in field tests. This agreement is reached thanks to the fact that the major contribution to the uncertainty calculation is the analysis reproducibility, determined during an inter-laboratory exercise. For the NOx automatic method, the comparison shows rather big discrepancies between both approaches and enlightens the conclusion that a rather good reliability can be reached using the uncertainty budget approach only if the following conditions are met: - trained personnel to correctly establish and interpret uncertainty budgets, - sufficient information on the performance characteristics of the method at the studied concentration, - knowledge of the variation range of influent parameters in the field, such as temperature, humidity, voltage, etc.