Invasive species have imposed huge negative impacts on worldwide aquatic ecosystems and are generally difficult or impossible to be eradicated once established. Consequently, it becomes particularly important to ascertain their invasion risk and its determinants since such information can help us formulate more effective preventive or management actions and direct these measures to those areas where they are truly needed so as to ease regulatory burdens. Here, we examined the global invasion risk and its determinants of sharpbelly (Hemiculter leucisculus), one freshwater fish which has a high invasive potential, by using species distribution models (SDMs) and a layer overlay method. Specifically, first an ensemble species distribution model and its basal models (developed from seven machine learning algorithms) were explored to forecast the global habitat-suitability and variables importance for this species, and then a global invasion risk map was created by combining habitat-suitability with a proxy for introduction likelihood (entailing propagule pressure and dispersal constraints) of exotic sharpbelly. The results revealed that (1) the ensemble model had the highest predictive power in forecasting sharpbelly's global habitat-suitability; (2) areas with high invasion risk by sharpbelly patchily spread over the world except Antarctica; and (3) the Human Influence Index (HII), rather than any of the bioclimatic variables, was the most important factor influencing sharpbelly' future invasion. Based on these results, the present study also attempted to propose a series of prevention and management strategies to eliminate or alleviate the adverse effects caused by this species' further expansion. (C) 2019 Elsevier B.V. All rights reserved.