During the Late Holocene, the majority of lowland river systems in temperate Europe transformed from low‐energy multi‐channel rivers in strongly vegetated marshy floodplains to more open floodplains characterised by single‐channel meandering rivers with overbank deposits. While the general framework of this transformation in floodplain geomorphology, ecology and hydrology (i.e. geoecohydrology) is widely recognised many uncertainties remain as its timing varies significantly, both among different river catchments and within them. To unravel whether the observed differences in floodplain response can be attributed to differences in the timing and nature of the driving forces or to a difference in sensitivity towards them, we compare long‐term and large‐scale reconstructions of the geoecohydrological floodplain dynamics and of the (climatically and anthropogenically driven) land cover change for two contrasting regions: the central Belgian loess belt and the sandy Campine region. By using a combination of cluster analysis, ordination and Ellenberg indicator scores on a large multi‐proxy and multi‐site dataset, we revealed the major trends in the past geoecohydrological evolution of northeastern Belgian floodplains. These trends are probably determined by changes in floodplain wetness, which can in turn be linked to variations in upland forest cover. The Early and Late Holocene floodplain transformations appear synchronous with the respective increases and decreases in upland forest cover in the vicinity of the sites, largely determining the water availability in the river catchments and thereby their local geoecohydrological conditions. Initially, these evolutions were determined by climate, but during the Middle and especially Late Holocene anthropogenic influence became a far more important factor, causing the evolutions in the two studied regions to increasingly diverge. While marshy floodplains with forested margins can still be found in the sandy Campine region today, these have become rare in the central Belgian loess belt due to the combination of a higher level of human impact and greater erodibility of the soils in this area. Despite the strong spatiotemporal variability of the floodplain transformations, we observe a trend towards increasingly rapid floodplain responses to upland land cover changes over the course of the Holocene, probably related to the growing hillslope–floodplain connectivity. We conclude that the (dis)similarities in Holocene geoecohydrological floodplain change can be largely attributed to the (dis)similarities in (climatically or anthropogenically driven) land cover change in the uplands, although the differences in inherent sensitivity of the locations – linked to factors such as soil type and topography – further complicate the already non‐linear impact–response relationships.