Understanding the particle dynamics is of paramount importance to the design, operation and optimization of looping combustion reactors. In this paper, a three-dimensional Eulerian-Lagrangian model was developed to study the particle flow in a dual circulating fluidized bed consisting of an air reactor and a fuel reactor. Based on the Multi-Phase Particle-In-Cell (MP-PIC) scheme, the turbulent gas motion was simulated by the large eddy simulation (LES) and the particle flow simulated by the discrete particle method, with the particle size distribution (d(p) = 0-108 mu m) being fully taken into account. The effects of particle size, particle density, initial bed inventory and fluidizing gas flow rate were numerically studied. The results show that small particles mainly distribute in the fuel reactor while large particles tended to stay in the air reactor, leading to an unfavorable influence on the efficiency of metal oxide. The increase of particle density led to the increase of global and internal circulation rates. A high initial bed inventory would result in a high particle concentration and lower axial velocities of particles. When increasing the fluidizing gas flow rate in the fuel reactor, the internal particle circulation rate increased gradually; however, the global particle circulation rate decreased. The results should be useful for the design and control of dual circulating fluidized bed, looping combustion reactors. (C) 2018 Elsevier Ltd. All rights reserved.