The use of NMR-derived chemical shifts in protein structure determination and prediction has received much attention, and, as such, many methods have been developed to predict protein chemical shifts from three-dimensional (3D) coordinates. In contrast, little attention has been paid to predicting chemical shifts from RNA coordinates. Using the random forest machine learning approach, we developed RAMSEY, which is capable of predicting both H-1 and protonated C-13 chemical shifts from RNA coordinates. In this report, we introduce RAMSEY, assess its accuracy, and demonstrate the sensitivity of RAMSEY-predicted chemical shifts to RNA 3D structure.