Abstract:Antarctic krill (Euphausia superba) is a key species in the Antarctic ecosystem. Investigating its habitat suitability can help the sustainable use of this resource and enhance our understanding of the Southern Ocean ecosystem. Such work is also useful for assessing krill population and exploring main fishing ground. However, different model algorithms may result in large deviations in calculating habitat suitability index (HSI) for krill resource. Therefore, this study aimed to explore appropriate methods for constructing HSI models for E. superba by using environmental factors such as sea surface temperature (SST), sea surface height (SSH), sea surface chlorophyll (SSC), and sea ice concentration (SIC). Two fitting methods, Neural Network (NN) model and Univariate Nonlinear (UN) model, were used to fit the environmental factors. Six algorithms, including minimum value, maximum value, product, arithmetic mean, geometric mean, and weighted arithmetic mean, were combined to construct HSI models. The results showed that the neural network model better predicted the actual distribution of habitats, while the univariate nonlinear regression method yielded more consistent results. The maximum and minimum value methods showed greater differences in their calculated results, introducing larger errors compared to other approaches. The continued multiplication method produced good predictive performance, while the arithmetic, geometric, and weighted arithmetic mean methods had similar results and were more stable. The proposed methods and conclusions in this study are of guiding significance for the assessment of similar species and the prediction of their habitats in the future.