Abstract: Based on the mesoscopic finite element model of particle reinforced composites , the effect s of stochastic properties of SiC particles on the ratcheting behavior of SiC_P/6061Al composites were numerically analyzed by employing a 3D multi-particulate unit cell and using an advanced cyclic plastic constitutive model. In the simulation, a 3D multi-particle unit cell containing the stochastic particle properties was first generated by the random sequential adsorption (RSA) method, and then the effects of the number of particles, the particle arrangement, shape, size and their stochastic distributions on the ratcheting of the composites were discussed by the numerical simulations. The results show that the smaller the particulate size and the higher the proportion of the particles distributing near the surface of matrix as well as the more the number of particulates contained in the unit cell, the higher the resistance to the ratcheting deformation. Meanwhile, the modeled composite with uniform distribution of particle size and location presents higher resistance to the ratcheting deformation than that with random dist ribution. The assumptions of spherical particle and its uniform distributions in size and location in the 3D unit cell can provide a reasonable simulation to the ratcheting of the particle reinforced metal matrix composites.