Doubly peaked density distribution is expected not only to affect the plasma-wetted area at divertor plates, but also to correlate with the upstream density profile and hence characteristics of magnetohydrodynamic activities in tokamak plasmas (Wang et al 2020 Phys. Rev. Lett.124 195002). Clarifying its origin is important to understand the compatibility between power/particle exhausts in divertor and high-performance core plasmas required by present-day and future tokamak devices. In this paper, we analyze the double-peak density profile appearing in the modeling during the physics design phase of the new lower tungsten divertor for EAST by using a comprehensive 2D SOLPS-ITER code package, including full drifts and currents, with a concentration on an unfavorable magnetic field (ion B × ∇B drift is directed away from the primary X-point). The results indicate that E × B drift induced by the plasma potential gradient near the target, which is closely related to the divertor state, plays essential roles in the formation of a double-peak profile at the target: (1) large enough radial Ep × B drift produces a broadened high-density region; (2) strong poloidal Er × B drift drives a significant particle sink and creates a valley on the high-density profile. Thus, the simulation results can explain why this kind of doubly peaked density profile is usually observed at the high-recycling divertor regime. In addition, features of the double-peak ion saturation current distribution measured in preliminary experiments testing the new lower tungsten divertor are qualitatively consistent with the simulations.