This paper mainly relates to the real polychlorinated biphenyl (PCB)-contaminated soil flushing process, in which an aqueous solution of anionic surfactant was passed through sandy soil having an average concentration 34.3 mg/kg of dry matter. The goal of the treatment was to decrease the PCB concentration in the soil to less than 10 mg/kg, which is a limiting value if the soil is to be used in the field of civil engineering. The laboratory part was focused on the demonstration of the suggested method for estimating the CMC value of the surfactant used for leaching the PCBs from the soil to the solution. The estimate was based on a set of batch experiments carried out with the same soil as the soil used for the pilot-scale experiments. Theoretically, all effects affecting the CMC should be considered in an estimated value. The experimental facility used for the pilot-scale demonstration consisted of a steel column (3 m in length, 1.5 m in diameter) containing 1.7 m(3) of polluted soil and a liquid circulation system, by which an aqueous solution of a surfactant was supplied to the soil. Spolapon AOS 146 (anionic surfactant) solution (40 g/L) was passing through the soil column for 2.5 months. The concentration of the surfactant and PCBs in the aqueous soil extract was monitored during this time period. The final PCB concentration profile in the soil was determined after stopping the liquid flow. After passing through the soil, the PCBs containing the aqueous extract was pumped out from the steel column bottom to a treatment unit, where it was processed by coagulation. The final PCB concentration profile in the soil was compared with the results of the theoretical model, which is also described in the paper. The time necessary to reach the limit demanded was estimated to move within the range from 6 to 12 months.