The purpose of the present study is to develop a new nanoparticle-classification technique to replace centrifugal separation method. To identify size-dependent particle properties, we investigated the dispersion, classification, and surface potential of silica particles treated with a bead mill, which was used to create a nanoparticle dispersion. The zeta potential of the particles in the dispersed slurry exhibited clear size dependence at tm = 30 min. That is, the zeta potential of the smaller particles was more negative than that of larger particles. Hence, the particles were classified using an EFFF system, in which the bottom and top plates of the channel served as negative and positive electrodes, respectively. Using a low applied potential, the 50% separation cut-size was 45 nm. This method prevented deposition of negatively charged silica particles on the bottom plate and allowed continuous operation. Results of theoretical calculations supported qualitatively the experimental results obtained in this study.