The random-fuzzy variables (RFVs) method based on the theory of evidence is studied, for the need of ADC uncertainty evaluation and the limitations of existing approaches. The connotation of RFVs adopted for expression of measurement result together with its associated uncertainty is discussed, and the RFVs mathematics for uncertainty propagation is analyzed. RFVs can naturally separate the contributions to the measurement uncertainty of the systematic and random effects. Taking power measurements as an example, RFVs method is applied to the presentation and propagation of the measurement uncertainty of ADC, and the results are compared with those obtained by GUM, which shows the RFVs method can be effectively employed in evaluating uncertainty of ADC, and is capable of providing the interval of confidence for all possible levels of confidence, within which the measurement result is supposed to lie.
