Abstract: Yellowfin tuna (Thunnus albacares) is one of the primary target species in China's tuna longline fishing industry. In order to improve fishing efficiency and promote sustainable use of resources, based on the longline production data of T. albacares in central Pacific Ocean from 2018 to 2022, combining with the environmental factors, such as temperature at different water layers, concentration of chlorophyll a, and salinity of sea surface (SSS), we analyzed the spatio-temporal distribution of the catch perunit of effort (CPUE) of T. albacares in that sea area and its relationship with environmental factors by constructing two models (GAM and nodeGAM). The results show that the peak season of T. albacares in the central Pacific Ocean was from May to July, and the CPUE was mainly distributed in the southern equatorial Pacific, with the highest tuna catches near 5°S. Compared with GAM Bias explained rate 29.4%, Mean-square error (MSE) 0.149, the nodeGAM showed significantly better goodness-of-fit with a bias explained rate of 49.4% (68.02% enhancement), and the MSE reduced to 0.103 (Decrease of 30.87%). Both models indicate that T₀ ≥27 °C, T₁₀₀ of 15−20 °C, SSS of 34.2‰−35.2‰, and CHL-a concentration of 0.1−0.25 mg·m⁻³ were the suitable environmental ranges for T. albacares. Compared with GAM, nodeGAM has a better fitting effect and explanatory ability. According to the effect curves of the model output, nodeGAM can better reflect the nonlinear relationship between CPUE and environmental factors.