The oceanic Nd budget is calculated using a steady state 10-box model and a compilation of field data. This is the first attempt to propose consistent estimates of the Nd fluxes entering the ocean, as well as indicating possible Nd sources and the proportion of Nd fluxes exchanged between dissolved and particulate fractions. With presently available Nd data the best estimates give a total Nd influx of 9 ¿ 109 g/yr, which leads to an oceanic Nd residence time of 500 years. From modeling tests we suggest that the authigenic Nd scavenged by particulates is 100% remineralized in the deep ocean. The total exchanged Nd flux may be as high as 2 ¿ 1010 g/yr. The ϵNd(0) values of the influxes are -22, -11, +1, and -4 for the North Atlantic, surface Atlantic, North Pacific, and surface Indo-Pacific regions, respectively. Atmospheric and riverine Nd fluxes are insufficient to explain the magnitude and regional variability of calculated Nd influxes and ϵNd(0). We propose continental margins as an additional source supplying Nd to the ocean. Using the model calibrated for Nd, we examine the sensitivity of deep water ϵNd(0) to variations of Nd inputs to the ocean. Deep water Nd concentrations and ϵNd(0) vary with the changes in Nd influxes and their ϵNd(0). As Nd sources to the ocean may change during glacial/interglacial periods, the ϵNd(0) shifts recorded in ferromanganese nodules and crusts do not necessarily reflect changes in paleoceanic circulation. The effects of continental erosion should be considered in reconstructing patterns of ocean circulation using Nd isotopes.