AbstractSwordtip squid (Uroteuthis edulis), which is sometimes eaten alive (lively squid) in northwest Kyushu, Japan, is an economically important fish species in the region. However, the total catch of this species in Japan has declined by more than 80% in the last three decades. To understand and predict the spatio‐temporal distribution of fish species, we developed a one‐dimensional ecosystem (NPZD) model and a habitat suitability index (HSI) model for southwest Iki Island, northwest Kyushu, Japan. Subsequently, we conducted three numerical experiments with the HSI model, with and without the NPZD model data (with the NPZD model data: phytoplankton or zooplankton concentrations, without the NPZD model data: only the physical data of the ocean). In the HSI model with zooplankton concentrations, we found a stronger positive relationship between the HSI model values and the daily fisheries catch data of U. edulis than that using only the physical variables of the ocean as the environmental parameters. Our study thus indicates that the performance of the fishing ground prediction model will improve by utilizing the lower trophic ecosystem model such as zooplankton concentrations. Furthermore, our results would provide important implications for the efficiency of fishing operations and the conservation and management of this species.