This work represents a first systematic approach to the size-based elemental quantification and size estimation of metal(loid) oxide nanoparticles such as silica (SiO2) in a real food matrix using asymmetric flow field-flow fractionation coupled online with inductively coupled plasma mass spectrometry (ICP-MS) and multi-angle light scattering (MALS) and offline with transmission electron microscopy (TEM) with energy-dispersive X-ray analysis (EDAX). Coffee creamer was selected as the model sample since it is known to contain silica as well as metal oxides such as titania at the milligramme per kilogramme levels. Optimisation of sample preparation conditions such as matrix-to-solvent ratio, defatting with organic solvents and sonication time that may affect nanoparticle size and size distribution in suspensions was investigated. Special attention was paid to the selection of conditions that minimise particle transformation during sample preparation and analysis. The coffee creamer matrix components were found to stabilise food grade SiO2 particles in comparison with water suspensions whilst no significant effect of defatting using hexane was found. The use of sample preparation procedures that mimic food cooking in real life was also investigated regarding their effect on particle size and particle size distribution of silica nanoparticles in the investigated food matrix; no significant effect of the water temperature ranging from ambient temperature to 60 °C was observed. Field-flow fractionation coupled to inductively coupled plasma-mass spectrometry (FFF-ICP-MS) analysis of extracts of both unspiked coffee creamer and coffee creamer spiked with food grade silicon dioxide, using different approaches for size estimation, enabled determination of SiO2 size-based speciation. Element-specific detection by ICP-MS and post-FFF calibration with elemental calibration standards was used to determine the elemental composition of size fractions separated online by FFF. Quantitative data on mass balance is provided for the size-based speciation of the investigated inorganic nano-objects in the complex matrix. The combination of FFF with offline fractionation by filtration and with detection by ICP-MS and TEM/EDAX has been proven essential to provide reliable information of nanoparticle size in the complex food matrix.