Precipitation is the key driver of terrestrial hydrologic cycle and the most important atmospheric input to land surface hydrology models. However, due to high elevation, complex terrain and inaccessibility, direct meteorological observations are either sparse or nonexistent in many remote parts of Tibetan Plateau (TP), posting a great challenge in hydrological modelling studies in the TP. This is especially the case for the Upper Brahmaputra (UB) basin in southearn TP. In this work, a gridded daily precipitation dataset with a spatial resolution of 10×10 km for 1961–2016 is reconstructed for the entire UB by combining orographic and linear corrections of China Meteorological Administration (CMA) and Global Land Data Assimilation Systems (GLDAS) data based on 262 rain gauges. The inverse evaluation of the reconstructed precipitation with a glacierhydrological model demonstrates the accuracy and rationality of the reconstructed precipitation for hydrological modeling in the UB. The reconstructed precipitation greatly improved the precipitation estimates in the UB with the basin-averaged mean annual precipitation increased from 465 mm in the original data to 709 mm in the reconstructed data for 1961–2016. The monsoon-dominated UB basin exhibits an overall negative precipitation vertical gradient with apparent orographic effects only appearing at local scales or in small basins. Basin size and climate control should be considered when using the orographic correction approach to derive high mountain precipitation on the Tibetan Plateau. The improved precipitation estimates for the UB may largely contribute to hydrological simulation and prediction studies in the basin.