Sea-level projections are highly anticipated outcomes of climate model simulations, relevant for coastal management worldwide. Ideally, any model simulation needs to be validated against observations, but this is impossible for the most recent sets of future climate model simulations, which start near to the present day (2020). Here, we compare satellite observations of regional sea-level change against projections from the Intergovernmental Panel on Climate Change Fifth Assessment Report for total sea-level change and its individual components over their overlapping period, from 2007–2022. We first test and compare three different methods to reduce the internal variability in the observations, which hampers the comparison with models, in particular for shorter time periods. While all three methods reduce the internal variability, we find the low-frequency component analysis (LFCA) removes most of the internal variability. We find that the regional projections are in good agreement with the LFCA-filtered observations, for 96% of ocean area within the 90% confidence interval. For the total sea level and sterodynamic component, the projections both under- and overestimate the observations, depending on the region. For mass-driven sea-level change, the regional projections tend to overestimate the observations. Our analysis gives confidence in sea-level projections for the instrumental era.