The organic pollutan adsorption/desorption process by microbial degradation had been less studied than metal ones. The sorption assays alone did not predict desorption, due to hysteresis, irreversibility, fixed compounds in different sites, with diverse desorption rates. Most of the studies dealt with bacteria rather than filamentous fungi and yeasts. So, our aims were to isolate yeasts from polluted sediments, to quantify its potential to uptake anthracene (An) and to evaluate the bioavailability by a desorption model. Yeasts were isolated from hydrocarbon-polluted samples, 40isolates grew in anthracene-plates. Molecular characterization was achieved by sequence analysis of the ITS1-5.8S rRNA-ITS4 and 26S rRNA regions; morphological and physiological determination were also done. Candida parasilopsis, Pichia anomala and Rhodothorula mucilaginosa were the prevalent yeasts. An-degradation was assessed in soilsystems with 0, 50, 100, 150, 200 and 250 μg An/l, 3 differentes sorbens types, organic carbon, organic nitrogen, PAHs, sand:silt:clay, pH and cation exchange capacity. Sophorolipids excretion were confirmed by HPLC, UV-detector with active fraction at 9.669 min (RT 9.646 min = sophorolipid-standard). A desorption model with equilibrium, nonequilibrium and nondesorption areas, was applied to explain the experimental data, An-transformation was greater in the organic liquid-phase than in the soil-sorbed ones; the desorption-coefficients and soil components were negatively correlated with the kinetic parameters. The An-release depended on the sophorolipid excretion, soil matrix and particles sizes. Desorption parameters significantly fitted the yeast uptake, with R 2 = 0.97, R 2 = 0.90 and R 2 = 0.97 for C. parasilopsis, P. anomala and R. mucilaginosa, respectively.