Accurate prediction of particle packing density remains challenging. Herein, the particle packing is simulated by discrete element modeling and the packing density results are used to evaluate the 3-parameter particle packing model. When the 3-parameter model is applied to binary and ternary blended mixes, the mean absolute errors in packing density prediction are 0.96% and 1.37%, respectively. Overall, the accuracy of the 3-parameter model is very good when applied to binary blends but not as good when applied to ternary blends, especially when the intermediate size particles are dominant. The discrete element modeling reveals that even when the intermediate size particles are dominant, there are significant particle interactions between the smaller size and larger size particles, which have not been properly accounted for in the existing packing models. Hence, there is still room for improving the existing packing models, requiring further discrete element modeling to study particle interactions.