We demonstrate low-coherence interferometry (LCI) for dye diffusion measurements in scattering tissue phantoms. The diffusion coefficient of a phthalocyanine dye in 1.5% agar gel containing scattering Intralipid was measured using a dual-wavelength interfero-meter. One wavelength was matched to the absorption peak of the dye at . The other, , was not affected by the dye, and was used to correct for varying sample scattering as a function of depth, assuming a constant ratio between scattering at the two wavelengths. The same wavelength dependence of scattering is assumed for the entire sample, but no a priori knowledge about the amount of scattering is needed. The dye diffusion coefficient was estimated by fitting a mathematical model of the interferometer signal to the measured LCI envelope. We compare results obtained using both a constant-scattering and a depth-resolved-scattering approach to determine the sample scattering. The presented method provides robust estimation of the diffusion coefficient when spatial resolution in determining the depth-resolved scattering is varied. Results indicate that the method is valid for samples having continuous spatial variations in the scattering coefficient over lengths as short as the coherence length of the probing light. The method allows in situ characterization of diffusion in scattering media.