In the Caribbean basin, particles less than or equal to 10 µm in diameter (PM10) have a significant epidemiological impact due mainly to the transport of desert dust. For the first time in this geographical area, the theoretical framework of artificial intelligence has been applied to predict PM10 concentrations. Six machine learning models were therefore used for the Guadeloupe archipelago. Gradient-boosting regression (GBR) is the model that gave us the best prediction performance. Comparing our results to other studies using machine learning models for megacities, we obtain similar performance using only three input variables while they use at least ten for their models. All these results show the specificities of PM10 concentrations in the Caribbean area.