Prediction of carbonate distributions at a global scale through geological time represents a challenging scientific issue, which is critical for carbonate reservoir studies and the understanding of past and future climate changes. Such prediction is even more challenging because no numerical spatial model allows for the prediction of shallow-water marine carbonates in the Modern. This study proposes to fill this gap by providing for the first time a global quantitative model based on the identification of carbonate factories and associated environmental affinities. The relationships among the four carbonate factories, i.e., "biochemical", "photozoan-T", "photo-C" and "heterozoan-C" factories, and sea-surface oceanographic parameters (i.e., temperature, salinity and marine primary productivity) is first studied using spatial analysis. The sea-surface temperature seasonality is shown to be the dominant steering parameter discriminating the carbonate factories. Then, spatial analysis is used to calibrate different carbonate factory functions that predict oceanic zones favorable to specific carbonate factories. Our model allows the mapping of the global distribution of modern carbonate factories with an 82% accuracy. This modeling framework represents a powerful tool that can be adapted and coupled to general circulation models to predict the spatial distribution of past and future shallow-water marine carbonates.