Estuaries, as highly dynamic marine environments, exhibit seasonal changes in fish habitat requirements, and the selection of appropriate analytical tools is critical to accurately reveal and understand such changes. Tree-based methods (including classification and regression tree (CART), random forest (RF), and conditional random forest (CRF)) have been widely used to explore the habitat requirements and predict the spatial distribution of marine species. However, the optimal methods still need to be comparatively evaluated based on the specific target species and region. This study focused on the important fishery resource Coilia nasus in the Yangtze River Estuary. Based on trawl monitoring survey data from the end of 2013–2018, seven explanatory variables, namely temperature, total nitrogen, latitude, year, salinity, pH, and ammonium, were used as predictors. The area under the receiver operating characteristic curve (AUC), Kappa, true skill statistic (TSS), and three other indicators were used to evaluate the overall performance of the three tree method models. The results showed that (1) almost all indicators pointed to the RF model having the best performance, while CRF and CART were significantly worse for some indicators, and the RF model had the best robustness and was suitable for habitat modeling in all seasons; (2) temperature and total nitrogen constituted the key habitat requirements of C. nasus, and habitat suitability rapidly increased or peaked in the temperature and total nitrogen ranges of 18.1–22.6 °C and 0–0.5 mg/L, respectively; (3) the spatial characteristics of C. nasus habitat in the Yangtze River Estuary exhibited seasonal changes in that the suitable area tended to expand gradually from spring to winter, and the northern branch of the Yangtze River and the offshore area of the Yangtze River were the main areas with high suitability for C. nasus. This study provides a useful reference for the spatial planning of the Yangtze River Estuary fishing ban area and the formulation of related conservation measures in the future.