Abstract: With the increasing application of geotechnical centrifuge model tests and the increasing demand of quantitative analysis, influencing factor and precise control of sample preparation on porosity and void distribution have arisen as a key issue in the development of geotechnical centrifuge modeling techniques. Sand pluviation is an important technique for preparing sand models and widely employed in geotechnical centrifuge modeling, due to the process of its reconstitution similar to the formation of natural coarse sand strata and the advantages of avoiding soil grain crushing and achieving a wide range of relative density. In order to understand the effect of control factors on the porosity of samples prepared by sand pluviation, a discrete element method based on PFC^3D is established to simulate the process of sample reconstitution using sand pluviaiton. Through discrete element simulation of the effect of drop height, mesh-grain size ratio, environmental medium, the feasibility of the proposed discrete element method is properly validated with the general knowledge from physical model tests. The results show that the porosity and the drop height appear in an obvious negative correlation, and the rate of decreasing in the porosity gradually attenuates with the increase of drop height. Meanwhile, a positive correlation exists between the mesh-grain size ratio and the porosity, and the latter is found to distinctly increase with the increasing of the former. The density of environmental medium is exhibited in a significant negative correlation with the porosity and has an important influence. Consequently, the layer thickness of model surface water plays a leading role in the control of the porosity.