The development of remote sensing of discharge algorithms has enabled us to estimate discharge without in-situ observations, yet their applicability of spatially continuous estimates of discharge is not adequately examined. This study calculated discharge for 668 continual reaches along the mainstem of the Yellow River using remotely sensed width from Landsat images and a discharge estimation algorithm based on the At-Many stations Hydraulic Geometry. Our result confirms the discharge decrease/increase at the dam/confluence points as well as the discharge reduction along the heavily irrigated areas, indicating the potential of capturing both natural and artificial discharge changes from the satellite. Moreover, the prediction fed by prior discharge, which gives the prior probability of the possible discharge values, that virtually considers the effect of water withdrawals improved the prediction accuracy, especially at downstream stations where the influence of irrigation is accumulated.