Predictions from species distribution models are utilized to parameterize the environmental factors that influence the biology, distribution, and habitats of a species of interest. This study fitted generalized additive models (GAMs) to spatiotemporal fishery data of light fishing from 2009 to 2013 to investigate the catch rates of swordtip squid in relation to changes in oceanographic conditions and developed a habitat preference model. A high Jensen–Shannon divergence (JSD) is considered to be an index of a thermal front. The results obtained using the selected GAMs revealed that the explained deviances in the catch rates pertaining to the oceanographic conditions was 45.10% throughout the year. All the variables used, the sea surface temperature (SST), chlorophyll-a, sea surface height anomaly, and JSD, were statistically significant predictors (p < 0.05), and the JSD explained the greatest amount of deviance (17.70%). The model predicted relatively high abundance of swordtip squid at 27–28°N in the southern East China Sea during spring and a decrease from June to August. The high abundance occurred again in September and extended to southwest ward to a region including coastal Mainland China. This demonstrated that the high abundance occurred in the 20.0–26.0°C SST range and 0.35–0.5 JSD range during spring around 27–28°N and movement towards southwest corresponding with shifts in the Kuroshio front (26°C isotherm) in summer and autumn.