Our understanding of African paleoclimate/hydrological history is mainly based on lake level and lake sediment studies. It improved during the last decade thanks to emerging stable isotope techniques such as compound-specific deuterium analysis of sedimentary leaf wax biomarkers (δ²H_leaf-wax). Here we present the results from a multi-proxy biomarker study carried out on a ∼100 ka loess-like paleosol sequence preserved in the Maundi crater at ∼2780 m a.s.l. on the southeastern slopes of Mt. Kilimanjaro in equatorial East Africa. The Maundi stable isotope records established for hemicellulose-derived sugars, lignin- and pectin-derived methoxyl groups, leaf wax-derived fatty acid and n-alkane biomarkers (δ¹⁸O_sugars, δ²H_{methoxyl groups}, δ²H_fatty-acids and δ²H_n-alkanes, respectively) reveal similar patterns, but also some distinct differences are obvious. The periods from ∼70 to 60 ka, the Last Glacial Maximum (LGM) and the Younger Dryas (YD) are characterized by more positive δ values, whereas during the Holocene, and around 30, 39, and 56 ka BP more negative δ values are determined. The application of a ‘coupled δ²H_n-alkane-δ¹⁸O_sugar paleohygrometer’ approach allows us to derive information about Late Quaternary changes of air relative humidity at the Maundi study site. Reconstructed changes of mean day-time relative humidity (RH_D) are in good agreement with pollen results from the study area. Apart from the overall regional moisture availability, the intensification versus weakening of the trade wind inversion, which affects the diurnal montane atmospheric circulation on the slopes of Mt. Kilimanjaro, is suggested as a local factor which may contribute to the observed variability of RH_D at Maundi study site. The combined usage of δ²H_n-alkanes and δ¹⁸O_sugars allowed us to reconstruct δ²H/δ¹⁸O of source water utilized by plants in the study area, which is directly linked to local precipitation. The results of this reconstruction caution against a straightforward interpretation of δ²H_leaf-wax and δ¹⁸O_sugars records as proxies for isotopic composition of local precipitation because variable and primarily RH-dependent isotopic evaporative enrichment of leaf water can mask changes of δ²H_prec/δ¹⁸O_prec in the past. The biomarker-based δ²H/δ¹⁸O_source-water records derived for the Maundi site revealed a discernible link with the reconstructed RH_D record; lower RH_D values were generally observed during periods characterized by more negative δ²H/δ¹⁸O_source-water values, indicating a reverse relationship with the expected precipitation amount. This indicates that the empirical relationship between amount of rainfall and its isotopic composition, observed nowadays on monthly timescale in the East African region, might not be valid on millennial time scale.