Since 1970, U.S. Army Corps of Engineers, Galveston District, has been using dredged sediments from the Houston ship channel to create and restore salt marshes in Galveston Bay. Some projects have failed due to excessive sediment erosion or siltation. The research reported here applies an engineering approach to analysis of tidal creeks in natural and beneficial use marshes of Galveston Bay. The hydrodynamic numerical model, DYNLET, was used to assess circulation in marsh channels. A preliminary sediment transport model was developed to analyze erosion and deposition for the same channels. In situ flume experiments were conducted to determine the sediment erodibility in natural and constructed marshes. A natural reference marsh, Elm Grove, was studied to understand marsh hydrodynamics and model calibration. The model results show that DYNLET can largely duplicate the marsh hydrodynamics and the sediment transport model can provide preliminary indication of erosion in tidal creeks. Analysis of the preliminary channel layout of the beneficial-use marsh demonstrated that channels will have sufficient circulation and optimum velocities.