Sea-level rise continues to be an issue of societal concern. It is a fact that millions of people live close to the coast and will be at risk both directly and indirectly from sea-level rise. Understanding the future begins in the past and so this thesis considers Holocene and modern sea-level variations in the Caribbean region, an area particularly at risk from long term sea-level rise. I construct a catalogue of 561 published Holocene sediment ages and depths, primarily fossil corals and mangrove peats. I derive probability distributions of the habitable ranges of corals and peat using modern growth and abundance records. These distributions are used to simulate realisations of sea-level position in the past. The relative sea-level (RSL) position and RSL rate are calculated at 500 year time slices to construct sub-regional sea-level histories. At each time slice, I select the realisations that fall within a 2000 year time window and calculate the least-squares estimate of RSL rate and RSL position. Results show Caribbean wide spatio-temporal RSL changes. From 7000 to 4000 cal yr BP, RSL rates were »2 mm yr−1 in the north (Cuba and Florida), »1 mm yr−1 in the east (Lesser Antilles and US Virgin Islands) and »2.5 mm yr−1 in the south west. From 4000 to 1000 cal yr BP, sea level rose between 3 (US Virgin Islands, Venezuela and Trinidad) and 5 metres (Florida, Belize). During the last 1000 years, RSL rates fell below 1 mm yr−1 and by 500 cal yr BP lay between 0 to 0.5 mm yr−1. The spatial variation between sub-regional RSL histories is also investigated by using a spherically symmetric, rotating numerical model that simulates sea-level change and vertical ground motion (VGM). I derive model RSL histories at the sub-regions in the Caribbean using a single deglaciation model and a range of earth parameters. By minimising the misfit of the model RSL curves to the data driven RSL curves, I find a representative model fit where lithospheric thickness is 71 km, upper and lower mantle viscosities are 0.5 and 10 × 1021 Pa s respectively. To find the change in sea-level rise in modern (1960 to 2012) times compared to the late Holocene, I calculate RSL and absolute sea-level (ASL) rates using 49 tide gauge records and satellite altimetry. I apply three corrections to remove seasonal and regionally coherent noise and calculate a least-squares estimate of sea-level change. Results show that present day RSL rise is up to three times greater than in the late Holocene, though there remains a small (· 1 mm yr−1) long term contribution from the last deglaciation. For individual sites, I found that variations in VGM can exacerbate long term sea-level change enhancing the risk of coastline communities.