Abstract. Barium is widely used as a proxy for dissolved nutrients and particulate organic carbon fluxes in seawater. However, these proxy applications are limited by insufficient knowledge of the dissolved distribution of Ba ([Ba]). For example, there is significant spatial variability in the Ba–Si relationship, and ocean chemistry may influence sedimentary Ba preservation. To help address these issues, we developed 4,095 models for predicting [Ba] using Gaussian Progress Regression Machine Learning. These models were trained to predict [Ba] from standard oceanographic observations using GEOTRACES data from the Arctic, Atlantic, Pacific, and Southern Oceans. Trained models were then validated by comparing predictions against withheld [Ba] data from the Indian Ocean. We find that a model using depth, T, S, [O2], [PO4], and [NO3] as predictors can accurately predict [Ba] in the Indian Ocean with a mean absolute percentage deviation of 6.3 %. We use this model to simulate [Ba] on a global basis using these same six predictors in the World Ocean Atlas. The resulting [Ba] distribution constrains the total Ba budget of the ocean to 122±8 × 1012 mol and clarifies the global relationship between dissolved Ba and Si. We also calculate the saturation state of seawater with respect to barite, revealing that the ocean below 1,000 m is, on average, at or near saturation. We describe a number of possible applications for our model output, ranging from use in biogeochemical models to paleoproxy calibration. Our approach could be extended to other trace elements with relatively minor adjustments and demonstrates the utility of machine learning to accurately simulate the distributions of tracers in the sea.