Low reducing agent operation of the blast furnace is an essential method for mitigating CO2 emissions in ironmaking. Because the coke rate is reduced in low reducing agent operation, gas permeability tends to deteriorate. Recently, blast furnaces with inner volume larger than 5000 m3 have become usual not only in Japan, but also in other Asian nations. Under these conditions, detailed information on in-furnace phenomena is required to attain stable operation.In the present study, a combination model using the discrete element method and computational fluid dynamics (DEM-CFD) was introduced to understand the fully three-dimensional in-furnace phenomena in the whole blast furnace. Due to the limitations of computational resources, the number of DEM particles must be reduced when applying DEM to the whole blast furnace. On the other hand, small cells must be used in the continuum model in order to calculate the gas flow in detail. Thus, mutual conversion between the location of particles in DEM and the property of the cells in the continuum model is needed. In this study, a method of converting information on the locations of cluster-approximated particles treated in DEM to continuum cells was proposed. Furthermore, optimization in which cells could be obtained without conversion parameters was performed to avoid losing local information obtained by DEM calculations. Simulations of solid movement and gas flow were successfully carried out with this coupled DEM-CFD model. As a result, it became possible to understand the three-dimensional stress field among particles under gas flow, transient gas flow and pressure distribution caused by charging of the burden materials, as well as solid motion.