Abstract: Flow rate and particle size are important environmental factors affecting arsenic migration and have received extensive attention in the study of arsenic migration mechanism. However, numerical modeling efforts addressing their combined effects remain limited. In this study, natural river sand was used as the porous medium, with As(V) solutions prepared for batch and column experiments to investigate kinetic adsorption, isotherm adsorption, and breakthrough behaviors. A numerical model was developed to simulate arsenic transport in river sand under varying flow velocities and grain sizes, and the controlling mechanisms were further examined through detailed characterization. Chemical adsorption is the primary mechanism of arsenic adsorption in river sand, occurring mainly at sites containing local aluminum and iron oxides, with both surface adsorption and intraparticle diffusion present. Fine-grained sand adsorbs more arsenic than coarse-grained sand, reaching adsorption equilibrium in a longer duration, and the breakthrough time of arsenic in the sand column significantly decreases with increasing grain size and flow velocity. The convection–dispersion–adsorption model effectively reproduces breakthrough experimental results (R²>0.97). Simulations further reveal that arsenic transport in coarse sand is more sensitive to flow velocity, with dispersion coefficients (D) increasing in both high-velocity fine sand and low-velocity coarse sand systems. Both the solid-liquid distribution coefficient K_d and the retardation coefficient R_d decrease with increasing flow velocity and grain size, further confirming that low flow velocities and fine sands are conducive to arsenic adsorption and adverse to its migration. This study can provide a scientific basis for modeling arsenic transport in groundwater systems.