Floods are considered as the costliest natural disasters. The nonlinear Muskingum models were widely applied to flood routing in a river reach with an inflow and outflow. In reality, many rivers have lateral inflows or are composed of several inflows (multiple inflows river) with nonlinear relation between the storage and inflow/outflow discharge. In this study, three variations on the nonlinear Muskingum model were extended to allow for lateral inflows and multiple inflows. The extension is based on an equivalent inflow for river systems with both lateral inflows and multiple inflows. A Genetic Algorithm (GA) was employed to estimate the extended nonlinear Muskingum models parameters. Two case studies (a river with a lateral inflow and multiple inflows) were considered to evaluate the extended nonlinear models. In addition to the sum of the squared deviation (SSQ), the efficiency of models was examined by the Nash-Sutcliffe efficiency (NSE). The sensitivity analysis of the extended models was performed to get more insight into the effect of the different values of models’ parameters on the objective function. The extended nonlinear Muskingum models outperform the multiple inflows linear Muskingum model and standard nonlinear Muskingum models in predicting the outflow hydrograph of a river system involving lateral inflows and multiple inflows. The outflow hydrographs of the case studies were precisely predicted with increased number of the extended models’ parameters in multiple inflows systems.