By analyzing the data collected over a rice paddy field in China, the occurrence of large drag coefficient in light winds (U ≤ 1ms −1) under near-neutral conditions (−0.1 ≤ z/L ≤ 0.1) is analyzed in this study. It is found that, with the use of either vector-or scalar-averaged wind speed, the drag coefficient under near-neutral conditions (C DN) always increases with the decrease of wind speed. Such behavior of C DN cannot be interpreted by the change of aerodynamic roughness length z 0 because it would need to be enlarged by an order of magnitude. While in light winds, the nonlocal transport, which is caused by large-scale eddies and overlooked by traditional Monin-Obukhov Similarity theory, is found to be the main cause of increased C DN. At last, a fixing term related to nonlocal effects is proposed based on our observation, and its applicability is verified with two AmeriFlux observation sites with a landscape of forest and savanna, respectively. Plain Language Summary According to the Monin-Obukhov Similarity (MOS) theory, the drag coefficient over land under near-neutral conditions (C DN) should be a constant and independent of wind speed because of the static surface roughness elements. However, the abnormal increase of C DN in light winds has been reported in many observations. In this study, we first examine the characteristics and mechanisms of the increased C DN in light winds. We conclude that the nonlocal transport caused by large-scale eddies is the major cause for the increased C DN in light winds. Furthermore, to successfully predict the variation of C DN with wind speed within MOS framework, we propose a fixing term related to nonlocal effects with our rice paddy observations, and the applicability of this fixing term is verified by two AmeriFlux observation sites with a landscape of forest and savanna, respectively. LIU ET AL.