Abstract: The vertical pressure at the bottom of the bulk grain pile in a large warehouse was measured and studied. The discrete element software PFC^3D was used to improve grain particles based on clump particle cluster element. The results of numerical simulation of bottom pressure were compared with field tests, and the distribution law of vertical pressure at the bottom of grain pile and its influencing factors were discussed. It can be concluded that the improved particle cluster element based on clump element can better simulate the non-uniformity of vertical pressure at the bottom of grain pile. A spatial distribution of the vertical floor pressure is presented along the whole bin, which is greatly different from that obtained from the recommended equation in the Code. The pressure increases with the distance from the corresponding bin wall. The particle shape, filling method and the friction between the wall and the grain affect the pressure. The coordination number of spherical particles and ellipsoid particles are 6.12 and 10.36, respectively. The ellipsoid particles are more prone to self-locking phenomenon, which can resist the fluctuation of vertical pressure at the bottom of grain pile. The vertical pressure at the bottom presented by ellipsoid particles has better dispersion and can better simulate the law of bottom pressure. The compaction state of grain pile structure is different due to the different forming modes of grain particles, which affects the distribution of vertical pressure at the bottom of the warehouse. It is shown that the friction between the bin and the granular grain, as well as the distance away from the bin wall, has a significant effect on the vertical floor pressure. The distribution law of vertical pressure at the bottom of the silo is different with different friction coefficient of the silo wall. A distribution coefficient of pressure is defined to quantify the pressure distribution of different locations in the bin bottom. The pressure distribution coefficient is small at the bin wall, whereas it gradually increases with the enhancement of the distance of corresponding wall surface. When it reaches a certain distance, the pressure distribution coefficient decreases slightly, but the decreasing range is small. As depicted in the investigation and examination, the distance away from the bin wall is smaller than 5 m, the distribution coefficient of pressure increased rapidly. After that, it will increase slowly until it becomes stable gradually at about 8 m away from the corresponding bin wall both in the results of field test and the discrete element modeling.