The ocean constitutes a minor source of atmospheric carbon monoxide (CO) at the global scale, but could play an important role far from continental anthropized emission zones. Here we use the NEMO-PISCES (Nucleus for European Modelling of the Ocean, Pelagic Interaction Scheme for Carbon and Ecosystem Studies) ocean general circulation and biogeochemistry model to dynamically assess the oceanic CO budget and its emission to the atmosphere at the global scale. The main bio-chemical sources and sinks of oceanic CO are explicitly represented in the model. The main processes driving the CO concentration are photoproduction and bacterial consumption and are estimated to 19.2 and 21.9 Tg C yr−1 respectively with our best-guess modelling setup. There are however very large uncertainties on their respective magnitude. Despite the scarcity of the in situ CO measurements in terms of spatio-temporal coverage, the proposed best simulation is able to represent most of the data (~ 300 points) within a factor of two. Overall, the global emissions of CO to the atmosphere are 3.6 Tg C yr−1, in the range of recent estimates, but very different from the ones published by Erickson in 1989, which were the only gridded global emission available to date. These oceanic CO emission maps are relevant for use by atmospheric chemical models, especially to study the oxidizing capacity of the atmosphere above the remote ocean.