The use of engineered nanoparticles (NPs) is increasing in many application fields, in fact nanomaterials are already commonly present in industrial, pharmacetutical, medicine, cosmetic, food products. While benefits of nanotechnology are widely publicized, the discussion of the potential effects of their widespread use in the consumer and industrial are just beginning to emerge, although the concern about the exposure to engineered nanomaterials is increasing. Inorganic oxides, such as titanium dioxide, TiO2, have been used by cosmetic industry for many years, even if their use is limited due to cosmetically unacceptable whitening effect imparted to the skin. TiO2 is widely used since it is considered insoluble, highly stable, non-reactive with other materials, thermally stable and nonflammable, low-cost and environmental friendly; however, it is also well known that either its chemical and physical properties and the adverse effects for man are size-dependent. For these reasons, there is a growing interest in finding techniques able to characterize nano-range particles by size. Flow FFF (FlFFF) is here proposed as separation technique to sort NPs because the separation is fast, occurs in a wet state and it is independent of particle density. The separation process is followed on-line with conventional UV-Vis detectors, even if specific detectors, such as ICP-AES and a polarographic analyzer, have been use to monitor the Ti on collected fractions during the FlFFF runs. In this work, TiO2 NPs are extracted from cosmetic formulations the solvent extraction procedures allowed to isolate TiO2 NPs from commercial samples, avoiding the use of large quantities of organic solvents, and leaving the NPs always in a wet state as required in the cosmetic formulations. The extracted TiO2 suspensions were stable so that their interaction with the FlFFF membrane (accumulation wall) during the separation was negligible. Consequently, the resulting particle size distributions (PSDs) derived by the FlFFF fractograms, should well represent the NPs in the cosmetic product when applied on human skin. Experiments using different cosmetic formulations, are presented and discussed to assess the large applicability of the proposed method; further work is already in progress in order to optimize the extraction and detection conditions.