Species Distribution Models (SDM) are important tools for exploring future changes in global biodiversity. However, relatively few studies have applied multi-model approaches to assess biogeographic changes globally, especially in marine environments. Antarctic krill (Euphausia superba) is a key species in the Southern Ocean marine ecosystem, but researches on its habitat are still limited, which hampers the understanding of its distribution and resource management. To this end, based on the resource density data of Antarctic krill and five environmental factors including Sea Surface Temperature, Sea Surface Chlorophyll, Sea Surface Height, Net Primary Productivity and Sea Ice Coverage, through ensembled models of 12 algorithms, we analyzed the habitat of Antarctic krill. Our results showed that, of the 8 algorithms, the Support Vector Machines, Random Forests, Multivariate Adaptive Regression Splines, Classification and Regression Trees, Multi-Layer Perceptron, and Recursive Partitioning and Regression Trees consistently outperformed the others; while the Bioclim, Domain, and Maxlike had poor performance, although they also differed among months. Besides, there was a link between inter-annual environmental variability and the spatial and temporal patterns of potential habitat for Antarctic krill. The results showed that the suitable habitat for Antarctic krill is mainly located in the central Bransfield Strait. The prediction of Antarctic krill habitat by ensembled model is helpful for the conservation of Antarctic krill resources, and for the exploration of the best fishing ground for Antarctic krill and the space of important fishing area in the future.