According to the fisheries data of Dosidicus gigas off Peru in the spring (August to October), summer (November to January), autumn (February to April) and winter (May to July) from 2006 to 2014, combined with three key marine environmental factors:sea surface temperature (SST), sea surface height (SSH), and net primary production (NPP), habitat suitability index models (HSI) with different weighting scenarios were employed to predict the fishing ground distribution of Dosidicus gigas off Peru. The relationship between catch per unite effort (CPUE) and SST as well as SSH and NPP was used to establish a suitability index (SI) model for each factor, and the arithmetic weighting method was applied to develop an integrated HSI model. Based on the proportions of catch and effort at each HSI class interval from the different weighted HSI models, the optimal models for different seasons were selected. The HSI model was validated by the data from 2015. The results showed that, the CPUE and the latitude gravity center of fishing ground (LATG) showed significant interannual and seasonal changes. In terms of interannual changes, CPUE basically fluctuated between 2 and 7 t/d, and the LATG basically ranged from 10°S to 18°S. On the seasonal change, CPUE in winter and spring was high, and the LATG was located in the northward regions on the fishing ground. In summer, the CPUE increased with the southward shift of the LATG. Compared with spring, the habitat moved 1.5° southwards and the CPUE increased by 6.7%. The change trend in autumn was the same as in winter and spring. The weighting schemes of the optimal habitat model in different seasons were different. The optimal weighting scheme of the spring was Case 9; the environmental factor with the highest weight was SST; the optimal weighting scheme of the summer was Case 7, and SST, SSH, and NPP weights were equivalent; the optimal weighting scheme for the autumn was Case 3, and the highest weighting was SSH. The above results indicated that the environmental factors had different impacts on the habitats of the Dosidicus gigas off Peru in each season. The accuracy of habitat predictions for each season was 84.68%, 78.56%, 72.74%, and 68.70%. It was helpful to understand the impacts of environmental factors on Dosidicus gigas stocks off Peru at different seasons, providing a scientific basis for the distant-water operations of Chinese squid-jigging fisheries.