The fractionation of natural nanoparticles by Asymmetrical Flow Field Flow Fractionation (As-Fl-FFF) was optimised by considering the following operating conditions: ionic strength, surfactant concentration and crossflow rate. The method performances such as fractionation recovery and fractionation efficiency were evaluated on a stable solution of colloidal-size natural inorganic particles. The online multi-detection by ultraviolet/visible spectrophotometer (UV) and multi-angle laser light scattering (MALLS) provided the monitoring of the sample during the separation and the evaluation of the fractionation efficiency. The lowest ionic strength and surfactant concentrations (i.e. 10(-3) mol L(-1) NH4NO3 and 3 x 10(-4) mol L(-1) SDS) allowed to obtain the highest sample recovery and lowest loss of the largest particles. The crossflow rate was investigated in order to avoid significant membrane-sample interaction. The applicability of the fractionation in optimised conditions was evaluated on a natural soil leachate, which was filtrated with different filter cut-offs. Filtration efficiency was stressed by the decrease of the large unfractionated particle influence in the void volume. For the first time, robust operating conditions were proposed to well size-fractionate and characterize soil nanoparticles within a single multi-detection analysis.