Stratigraphic data from salt marshes provide accurate reconstructions of Holocene relative sea-level (RSL) change and necessary constraints to models of glacial isostatic adjustment (GIA), which is the dominant cause of Late-Holocene RSL rise along the U.S. mid-Atlantic coast. Here, we produce a new Mid- to Late-Holocene RSL record from a salt marsh bordering Great Bay in southern New Jersey using basal peats. We use a multi-proxy approach (foraminifera and geochemistry) to identify the indicative meaning of the basal peats and produce sea-level index points (SLIPs) that include a vertical uncertainty for tidal range change and sediment compaction and a temporal uncertainty based on high precision Accelerator Mass Spectrometry radiocarbon dating of salt-marsh plant macrofossils. The 14 basal SLIPs range from 1211 ± 56 years BP to 4414 ± 112 years BP, which we combine with published RSL data from southern New Jersey and use with a spatiotemporal statistical model to show that RSL rose 8.6 m at an average rate of 1.7 ± 0.1 mm/year (1σ) from 5000 years BP to present. We compare the RSL changes with an ensemble of 1D (laterally homogenous) and site-specific 3D (laterally heterogeneous) GIA models, which tend to overestimate the magnitude of RSL rise over the last 5000 years. The continued discrepancy between RSL data and GIA models highlights the importance of using a wide array of ice model and viscosity model parameters to more precisely fit site-specific RSL data along the U.S. mid-Atlantic coast.