The semi-distributed, hydrological HBV model is commonly used in hydrological watershed simulation; however, its simulation performance varies significantly across different watersheds. Identifying key factors influencing the simulation performance can help improve the application accuracy of HBV model in different watersheds. This study employed the HBV model to simulate the daily flow of 38 sub-basins in the Yangtze River Basin. The determination coefficient (R2), Nash-Sutcliffe efficiency coefficient (NSE) and relative error (RE) were analyzed, and the impact of 12 factors, including land use, topography, and meteorological hydrology, on the simulation accuracy was explored using a random forest model. The results showed that the HBV model was highly applicable to the Yangtze River Basin, with average R2, NSE, and RE values of 0.62, 0.56 and -0.19 for the 38 sub-basins, respectively; however, the simulation accuracy of the HBV model varied significantly among the different sub-basins. Furthermore, the size of each evaluation index showed random spatial distribution, indicating that the simulation accuracy of the HBV model was affected by non-systematic factors. Among the influencing factors, the density of rainfall stations was the most significant factor affecting model simulation accuracy, followed by average annual temperature and precipitation; whereas the impact of topographical features was relatively low compared to meteorological and hydrological factors. The HBV model demonstrated a notable ability to achieve high simulation accuracy in sub-basins with various spatial scales and meteorological terrains in the Yangtze River Basin. With the continuous development of high-precision precipitation products, the HBV model has broad application prospects in hydrological simulation and water resources management in China.