The rhizosphere is generally depleted in nutrients, but as a hotspot of microbial activity it fosters crop P uptake. We hypothesized that P contents of water extractable nanoparticles (<0.1 μm) and small sized colloids (<0.45 μm) differ between non-rhizosphere and rhizosphere soil. To test this hypothesis, rhizosphere and non-rhizosphere soils (Luvisol and Cambisol) were sampled at harvest period of winter wheat near Selhausen (Germany). Microaggregate and colloidal fractions in the size range of 53–250 μm, 20–53 μm, 0.45–20 μm, and <0.45 μm were separated by wet-sieving and centrifugation. Subsequently, the colloids <0.45 μm were further isolated in 0.66–20 nm, 20–100 nm and 100–450 nm fractions using asymmetric flow field flow fractionation (AF4) and directly analyzed by online coupled organic carbon detector (OCD) and inductively coupled plasma mass spectrometry (ICP-MS) for element composition. No significant differences (p > 0.05) were measured between rhizosphere and non-rhizosphere soil P contents of microaggregate fractions. The rhizosphere soil, however, showed ∼26 % depletion of average P content in the 0.66–20 nm fraction, which went along with an enrichment of P content of the 100–450 nm fraction by a factor of two. Apparently, P uptake by plants results in a redistribution of P in the rhizosphere, with small nanoparticles providing available P to plants while excess residual P is bound to fine colloids.