The impact of small-scale fading on the estimation of local average power density for radiofrequency exposure assessment is studied in the case of a Rayleigh fading, a Rician-

fading, and a Nakagami-

fading. In all cases, it is shown that the significant parameter is the coefficient of variation of power density. We give the relation between the error on the estimation of local average power density and the number of independent points that are used in the averaging process. We apply this analysis to a distribution of fading statistical properties provided by measurements.