Dielectrophoresis has proven to be an effective method for the separation of bioparticles such as cells. Nevertheless, the electric polarization induced by nonuniform electric fields leads to a dipole-dipole interaction between particles and therefore the formation of chains is likely to occur. In this paper, we will present an approach based on a drift-diffusion dynamics to quantitatively study formation and kinetics of particle-chains via the introduction of the particle stitching as chemical-like reactions. This approach will allow us to dynamically describe, in the framework of a numerical simulation, particle clustering, thus providing a suitable tool for reproducing data from dielectrophoretic experimental setup.