Understanding loess sedimentation rates is crucial for constraining past atmospheric dust dynamics, regional climatic change and local depositional environments. However, the derivation of loess sedimentation rates is complicated by the lack of available methods for independent calculation; this limits interpretation of the environmental changes revealed by the loess record. In particular, while the Quaternary/Neogene Chinese loess and Red Clay sequences have the potential to provide detailed records of past sedimentation and climate change, there is great uncertainty concerning: (i) the influences on sediment grain-size and accumulation; and (ii) their relationship through time and across the depositional region. This uncertainty has led to the widespread use of assumptions concerning the relationship between sedimentation rate and grain-size in order to derive age models and climate reconstructions. To address this uncertainty, detailed independent age models, based on optically stimulated luminescence dating, undertaken at 10 to 40 cm intervals at five sections across the Loess Plateau in China, have been used to calculate sedimentation rates and make comparisons with grain-size changes over the late Pleistocene and Holocene. The results demonstrate that sedimentation rates are site specific, extremely variable over millennial timescales and that this variation is often not reflected in grain-size changes. In the central part of the Loess Plateau, the relationship between grain-size and sedimentation rate appears most complex, suggesting an interplay between local conditions at source and sink and a changing emplacement mechanism. This observation further undermines the common use of loess sedimentation age models that rely on a derived relationship between grain-size and sedimentation rate from a type section. The results also highlight the difficulty in assigning specific environmental causes to sedimentation rate changes and, to a lesser extent, grain-size shifts.