Abstract: Abstract: Near-bed boundary conditions are of great importance to numerical simulations of sediment laden flows in open channels. The experiments observed that the velocity distributions of hyper-concentrated sediment-laden flows differed significantly from that of clear water flows, due to the turbulence damping caused by suspended sediment. With the increase of sediment concentration in sediment-laden flow, the turbulence intensity decreased, while the velocity decreased near the channel bed and increased near the flow surface. The near-bed boundary condition for clear-water flow, without considering the turbulence damping from the interactions between two phases, generated errors in the simulations of the hyper-concentrated sediment-laden flow. In this study, a mixture model derived from the Eulerian two-phase model was employed, which maintained the computational efficiency of a simple diffusion model and the similar accuracy with the Eulerian two-phase model. Newly developed relative velocity formulations were employed to perform the interactions between the sediment and water. The km-εm model was applied to close the Reynolds stress of the mixture phase, which performed the turbulence damping effect well without the empirical constant. The mixture model was extended to the cases with a wide range of sediment sizes using the sediment diffusion enhancement from the particle wake. Five near-bed boundary conditions were selected to apply for the mixture model. These boundary conditions included the Dirichlet Boundary Condition (DBC) imposing boundary velocity, Neumann Boundary Condition (NBC) imposing boundary velocity gradient, Standard Wall Function (SWF), and two Modified Neumann Boundary Conditions (MNBC-ME and MNBC-MK). MNBC-MK and MNBC-ME respectively modified the turbulent kinetic energy and the dissipation rate of turbulent kinetic energy based on NBC. The five near-bed boundary conditions were tested in the simulations of the classic experiments of the sediment-laden flows in open channels. Their performances were then evaluated within the distributions of velocity, turbulent kinetic energy, eddy viscosity, and sediment concentration. Specifically, the computed velocity and sediment concentration of the DBC, SWF, and MNBC-MK were the better agreement with the experimental data in the dilute and hyper-concentrated cases. The velocity at near-bed boundary was given using the velocity log law in DBC, where the velocity parameters were adjusted to fit the experimental velocity in the hyper-concentrated cases. SWF presented the excellent self-adaptive ability to automatically regulate the velocity and turbulent kinetic energy under the cases with different sediment concentrations. NBC performed well in the simulations of the dilute cases, but failed to predict the velocity and sediment concentration in the hyper-concentrated cases. In the hyper-concentrated cases, the larger eddy viscosity was obtained by NBC and responsible for the underestimated velocity in the near-bed region and the overestimated one in the upper region, and also the overestimated sediment concentration. Compared with the NBC, MNBC-ME indicated no improvement on the computations, though it increased the dissipation rate of turbulent kinetic energy at the near-bed boundary in the hyper-concentrated cases. MNBC-MK introduced an empirical coefficient to modify the turbulent kinetic energy at the near-bed boundary, where the coefficient tended to decrease with the increase of sediment concentration. The coefficient calibration is required in DBC and MNBC-MK and limits the applicability in complex engineering. In a word, SWF is strongly recommended for the mixture model, in terms of the convenience and accuracy in simulating the engineering sediment-laden flow, such as the water diversion of the Yellow River and the sediment deposition in pump station.