In cooperation with the City of Wichita Falls, the U.S. Geological Survey assessed channel changes on the Wichita River at Wichita Falls, Texas, and modeled historical floods to investigate possible causes and potential mitigation alternatives to higher flood stages in recent (2007 and 2008) floods. Extreme flooding occurred on the Wichita River on June 30, 2007, inundating 167 homes in Wichita Falls. Although a record flood stage was reached in June 2007, the peak discharge was much less than some historical floods at Wichita Falls. Streamflow and stage data from two gages on the Wichita River and one on Holliday Creek were used to assess the interaction of the two streams. Changes in the Wichita River channel were evaluated using historical aerial and ground photography, comparison of recent and historical cross sections, and comparison of channel roughness coefficients with those from earlier studies. The floods of 2007 and 2008 were modeled using a one-dimensional step-backwater model. Calibrated channel roughness was larger for the 2007 flood compared to the 2008 flood, and the 2007 flood peaked about 4 feet higher than the 2008 flood. Calibration of the 1941 flood yielded a channel roughness coefficient (Manning's n) of 0.030, which represents a fairly clean natural channel. The step-backwater model was also used to evaluate the following potential mitigation alternatives: (1) increasing the capacity of the bypass channel near River Road in Wichita Falls, Texas; (2) removal of obstructions near the Scott Avenue and Martin Luther King Junior Boulevard bridges in Wichita Falls, Texas; (3) widening of aggraded channel banks in the reach between Martin Luther King Junior Boulevard and River Road; and (4) reducing channel bank and overbank roughness. Reductions in water-surface elevations ranged from 0.1 foot to as much as 3.0 feet for the different mitigation alternatives. The effects of implementing a combination of different flood-mitigation alternatives were not investigated.