Tests of degradation in a photocatalytic membrane system with the lamp immersed in the suspension inside the photoreactor have been carried out by using polycrystalline TiO2 (Degussa P25) as catalyst and humic acids, organic dyes, 4-nitrophenol as pollutants. The influence of the type of nanofiltration membrane, initial concentration of pollutant and pH on the photodegradation rate was investigated in discontinuous and continuous configurations. Two membranes were tested, i.e. NF-PES-010 (Celgard, Germany) of polyethersulphone and NTR-7410 (Nitto-Denko, Japan) of sulphonated polyethersulphone. The last one was chosen for all of the photoreactivity experiments because permeability and rejection tests indicated that it was able to hold both catalyst and small molecules carrying the same membrane charge (negative), thanks to the Donnan Exclusion. Despite the fluxes ranged between 20 and 401/(h m(2)) in operating conditions at 6 bar and these values are interesting for application purposes, the rejections of NTR-7410 nanofiltration membrane, obtained during operation of the membrane photoreactor in the degradation of humic acids, patent blue dye and 4-nitrophenol, were significantly lower than those obtained in the absence of photodegradation probably because of the small molecular size of by-products and intermediate species generated during the photodegradation process. This means that in order to select a suitable membrane, rejection should be determined during operation of the photoreactor.