Storm-surge models are commonly used to assess the impacts of hurricanes and coastal storms in coastal areas. Including the impact of the projected future sea level rise (SLR) in these models is a necessary step for a realistic flood risk assessment. Commonly, SLR is superimposed linearly on the estimated water elevation. This approach, while efficient, may lead to inaccuracies. Further, developing a new model with updated data that include the impacts of SLR (i.e., nonlinear approach) is time consuming. We compare the linear and nonlinear approaches to include the effect of SLR to predict Maximum Water/Flood Elevations (MWE) as a result of storm surge and SLR. After a simplified theoretical analysis, a number of idealized cases based on the typical coastal bodies of water are modeled to assess the impact of SLR on MWE using the linear superposition and nonlinear approaches. Additionally, two case studies are carried out: Narragansett Bay, RI and Long Island Sound, CT (USA). Results show that for the idealized cases with variable depth, in general, the linear superposition of SLR to MWE is conservative (i.e., predicts a larger flood elevation) relative to the nonlinear approach. However, if a constant depth is considered, results are not consistent (i.e. linear superposition can overestimate or underestimate MWE, and the results depended on the geometry). The simulated MWE from the Narragansett Bay simulation confirms the outcome of idealized cases showing that linear assumption is conservative up to 10% relative to the nonlinear approach. For this study, Hurricane Sandy and a Synthetic Storm from {the US Army Corps of Engineers} North Atlantic Comprehensive Coastal Study (NACCS) dataset are simulated. Long Island Sound model results are also consistent with the idealized case. In general, based on the results of the idealized and real studies, a discrepancy of up to 10% between the linear and nonlinear approaches is expected in estimation of MWE which can be under- or over-estimation of flood elevation.