Albacore (Thunnus alalunga) is a migratory pelagic species, and its spatio-temporal distribution is vulnerable to environmental variation. Thus, a better understanding of the environmental effects on the albacore habitat is of great scientific importance. We used information from the albacore logbooks of mainland China commercial longline vessels and the oceanographic environmental data in the area of 140°E-130°W, 0°-50°S for the South Pacific fishing season (from May to August) from 2015 to 2017 to analyze the response curves of the environmental factors affecting albacore catch per unit effort (CPUE) and the contribution rate of the environ­mental factors through a maximum entropy model (MaxEnt). We also explored the potential albacore habitat in the main 2017 fishing season and assessed the prediction accuracies compared to the actual catch data. The results showed that:(1) the optimal range of environmental factors were homologous:28.4-30.6℃ of sea surface temperature, 13.2-17.6℃ of sea temperature at 300 m depth, 35.6-36.7 of sea surface salinity, -1.6-5.8 m/s of northward sea surface wind north of 25°N, and 17.8-23.4℃ of sea surface temperature, 12.2-16.9℃ of sea temperature at 300 m depth, 35.2-36.0 of sea surface salinity, -0.7-4.9 m/s of northward sea surface wind north of 25°S. (2) The environmental factors (sorted by importance) north of 25°S were sea surface temperature (31.3%), sea temperature at 300 m depth (30.1%), sea surface salinity (29.2%), and northward sea surface wind (9.4%); north of 25°S, the environmental factors were sea surface temperature(60.7%), sea surface salinity (22.4%), northward sea surface wind (10.6%), and sea temperature at 300 m depth (6.3%). The most important environmental factor north of 25°S was sea surface temperature (over 60%, on average), which was more significant than the same variable north of 25°S. The importance of the first three environmental factors was similar north of 25°S (approximately 30%, on average). (3) The overall prediction accuracy was 30%-85%; the prediction accuracy of the medium potential habitat was relatively high, while prediction accuracies for the high and low potential habitats were low, as a result of the model and limited data.