Abstract: The water level-discharge relationship at Luoshan Hydrological Station is complex, which poses challenges for accurate water level forecasting.Enhancing forecast accuracy and extending the effective forecast period are pressing issues that need to be addressed.Based on the measured water level and discharge data and the upstream and downstream control node data from 2010 to August 2024, Spearman correlation analysis, support vector regression and water balance analysis method were used to systematically analyze the key influencing factors for the water level-discharge relationship at Luoshan Station.Based on the analysis results, a water level forecasting model was constructed.Results indicate that the key influencing factors of the water level-discharge relationship at Luoshan Station include simultaneous water level changes between Hankou Station and Luoshan Station, and the discharge from Lushui Reservoir in the previous day.Additionally, factors affecting water level forecasts also involve inflows from upstream tributaries and rainfall in Dongting Lake area in the previous day.The forecasting model demonstrates high accuracy during both the training period and the test period in the 1 to 7-day forecast period, and the R-squared is all above 0.97, indicating that the model can be applied in the actual forecasting at Luoshan Station.The research outcomes provide robust technical support for flood control operations in the Chenglingji area, while also offering a methodological approach for hydrological forecasting in complex river network areas.