Sediment transported by rivers is a major cause of concern as it may create harmful effects to life and structure. Therefore the amount of sediment transported through river system needs to be forecasted for its proper monitoring and control. Storage parameter influences the sediment flow. Hence various types of models incorporating sediment storage within the models needs to be investigated. The storage can be split into different parts as per the number of gauging sections in the river system, known as fractional storage. Adding up all the fractional storages lead to total storage in the river system. The variation of fractional storage and total storage has been studied for streamflow forecasting models previously. However, the variation of sediment storage in sediment models needs to be studied. In the present study, Multiple Input-Multiple Output (MIMO-1 and MIMO-2) and Multiple Input Single Output (MISO) models have been developed based upon complementary Muskingum equations for one day ahead concurrent sediment discharge prediction at the Mississippi river system, U.S. Normalized root mean square error (NRMSE) and Nash Sutcliffe coefficient of efficiency (CE) was acceptable (NRMSE < 10% and CE > 0.8) for most of the outputs. MIMO-2 model considering total storage variation performed marginally better than other models in terms of these statistical criteria. Self-certification of MIMO-1 models makes it more efficient for issuing real-time sediment flow forecast in river systems.