Abstract: Predicting a water surface elevation (WSElev) at a particular location has a wide range of applications like determining if a levee will overtop or how much a dike notch will increase water flow into a secondary channel. Five existing methods for predicting the water’s surface, (1) daily slope, (2) average slope, (3) River Analysis System (RAS) 1D, (4) RAS 2D, and (5) Adaptive Hydraulics modeling system (AdH), were used to predict the Mississippi River’s daily water surface from 10 October 2014 to 31 May 2016 at Friar’s Point, Greenville, and Natchez gages. The error, calculated as the model-predicted water surface minus the gage-observed water surface, was compared among the methods. The average slope method, using Helena and Fair Landing gages, and the daily slope method, using either Memphis and Helena or Helena and Arkansas City gages, most closely estimated the observed WSElev. The RAS 1D predictions for Friar Point and Greenville produced more accurate estimates than the RAS 2D model and were the only estimates that did not show a pattern of over- or underestimation. When the daily slope method was applied to gages that were farther apart (Memphis and Arkansas City, Arkansas City and Vicksburg, or Vicksburg and Knoxville), the error became greater than most RAS 1D and 2D predictions. The low error and simple calculations of the daily slope and average slope methods using gages <110 river miles apart make these methods useful for calculating current and historic conditions. The lack of over- or underestimation in the RAS 1D predictions (for locations away from the edges of the model area) make this method a better choice for predicting average WSElevs and a good choice for forecasting future WSElevs.