Packing density models could predict the bulk density and void ratio of particulate materials. The original linear packing density model (LPDM) is linearly related to each grain size fraction. However, when using the LPDM model, the predicted packing density and void ratio have large deviation compared with the experimentally measured values. The authors believe this is due to the wedging effect. Therefore, the LPDM model is modified based on wedging effect and converted into a nonlinear model. In this paper, the novel nonlinear model is extended to ternary mixtures. Then, the wedging effect parameter is fitted through the experimental data of binary mixing of quartz sand particles. The obtained wedging effect parameter is correlated to the characteristic diameter ratio between particles. The novel nonlinear model is verified via measured values and cited data from binary and ternary mixes. The results show the calculated values of the novel nonlinear model are closer to the experimental results than that of LPDM.