Poplars (Populus alba, Populus canadensis, Populus canescens, Populus deltoides, Populus fremontii, Populus nigra and Populus simonii) are found throughout the world and are invasive in South Africa, where they are spatially permitted in certain areas under controlled conditions, as specified in the country’s invasive species legislation. To better trace their geographic distribution, this study predicts the potentially suitable habitat of poplar trees in South Africa based on generalised linear model (GLM), Random Forests (RF) and Support Vector Machines (SVM) models and also assesses the climatic variables with the greatest impact on prediction performance. The results show excellent performance for all models (Area Under the Receiver Operation Characteristics Curve [AUC] 0.9) in predicting the poplar distribution, with RF achieving the best performance (r = 0.83 and AUC = 0.965), followed by SVM (r = 0.72 and AUC = 0.959) and then GLM (r = 0.65 and AUC = 0.937). In a geographical perspective, all models show a similar pattern, with the highest concentrations being in the south-western parts of the Western Cape, the Southern Cape on the Garden Route, the central-eastern Free State, Mpumalanga, and the southern parts of Limpopo. The evaluation of the relative importance of the bioclimates used showed that the warmest and driest quarter’s precipitation and annual precipitation significantly contribute to the poplar population. These results demonstrate the power of machine learning and regression models for predicting suitable habitats and extracting valuable environmental-climatic knowledge for monitoring and managing invasive tree species such as poplars.Conservation implications: Poplars are among the most aggressive invasive plant species in South Africa. The results of this study are expected to help conservation authorities understand the current climatic factors affecting the species distribution, as well as potential sites.