1D hydraulic models are largely used to simulate the propagation of flood waves and for flood mapping along river systems. The most common approach to account for the hydraulic effects of vegetated floodplains consists of imposing higher roughness coefficients. However, the flow resistance of vegetation is governed by plant submergence, which is water-depth dependent and varies with the discharge, and thus with time. An improved method properly incorporating floodplain vegetation roughness in 1D models is presented here. The Manning coefficient is derived from a simplification of Baptist’s formula assuming horizontal floodplains, i.e., with the same water depth everywhere. Considering the dependency of vegetation roughness on local water depth (in case of variable flow conditions), a predictor-corrector approach of the derived formula is proposed to be applied at every computational time-step. If different types of vegetation are present, the roughness coefficient, one for each floodplain, is derived as a weighted average. The method is tested on a recently restored stream located in the Netherlands, the Lunterse Beek, using the HEC-RAS code. The results support the implementation of the proposed method, but validation is needed for river floodplains with non-uniform vegetation cover.