For marine fish and invertebrates, larval dispersal plays a critical role in determining connections among source and sink habitats, and the lack of a predictive understanding of larval dispersal is a fundamental obstacle to the development of spatially explicit restoration plans for marine populations. We investigated larval dispersal patterns of eastern oyster in an estuary along the Northern Gulf of Mexico under different simulation scenarios of tidal amplitude and phase, river discharge, wind direction, and larval vertical migration, using a coupled biophysical transport model. We focused on the dispersal of larvae released from the commercially exploited (Cedar Point, CP) and non-exploited (Bon Secour Bay, BSB) oyster populations. We found that high flushing rates through the dominant inlet prevented larval exchange between the commercially exploited and non-exploited populations, resulting in negligible connectivity between them. Variations in tidal amplitude, river discharge and wind direction played a more important role in the amount of larvae retained in Mobile Bay when they are released from CP than from BSB. Under most of the scenarios, larvae from BSB were retained around the spawning area, while larvae from CP showed a predominant westward flow. Net sinking behavior of late-stage larvae increased larval retention in the bay, but physical transport showed a higher impact in the amount of larvae retained. These findings have enhanced our understanding of larval dispersal of eastern oyster in a wide, shallow estuarine system, and been used to establish spatially explicit strategies for oyster restoration in the Mobile Bay system, Alabama.