The green sand molds, which are used to manufacture cast iron castings, are made by compacting green molding sand particles with mechanical forces and must have appropriate mold strength and voids. Y. Maeda et. al.[1][2] are currently developing the simulation of the green sand molding process using Discrete Element Method (DEM), which is a numerical analysis method based on the Lagrangian system developed by P.A. Cundall et al.[3]. The simulation has the advantage that it can take into consideration the grain size distribution of green sand, but there is insufficient consideration of how to handle the particle size distribution. Actual test sands are granulated in nature and have a variety of grain sizes and shapes, and it is no exaggeration to say that no two are alike. The particle size distribution in nature is considered to follow a logarithmic normal distribution. Here, the particle size distribution of natural sand is referred to as a stepless particle size distribution. On the other hand, the production management of green sand properties uses the actual values measured through sieving at several steps. The particle size distribution in DEM is handled using production management values. Therefore, it is a clear difference between the two. In the present study, in order to handle the step-less particle size distribution, we attempted to estimate the probability density function by using a lognormal distribution based on measurements of green sand divided by several meshes. The several green molding sands, including silica sand, could be adequately represented by a log-normal distribution of numbers ratio. Also, the particle size distribution made from the probability density function is close to the lognormal distribution as increases as the division number. Further it was clear that the division number of particle size distribution influences sand compacting behavior on DEM simulation.