Abstract: Based on the water quality monitoring data of the main drainage ditch of the Wuliangsuhai Lake from 2011 to 2020,the patterns and characteristics of water quality changes were deeply investigared by employing statistical methods,the comprehensive pollution index method and principal component analysis,then a water quality prediction model was established using multivariate stepwise linear regression method to screen key indicators and optimize water quality monitoring process.The results indicated that from 2011 to 2020,the main monitoring indicators in the main drainage ditch,including biochemical oxygen demand,chemical oxygen demand,ammonia nitrogen,total phosphorus and total nitrogen,all showed a significant downward trend.By 2020,all monitoring sections reached the class V standard of surface water,with some indicators (ammonia nitrogen and total phosphorus) reaching the class Ⅱ standard.The result of comprehensive pollution index evaluation showed that the overall water quality of the drainage ditch had improved from a heavily polluted to a clean level,with water conditions at the main monitoring sections becoming stable and consistent.The results based on parameter optimization methods showed that the multivariable regression equation composed of TN,BOD₅ and TP could accurately predict comprehensive water quality changes.According to the analysis of major pollutants,the water quality of the main drain ditch flowing into the Wuliangsuhai Lake was mainly affected by both natural conditions and human activities,showing ecological risk characteristics of high mineralization and eutrophication.This provides solid data support and systematic theoretical foundation for subsequent lake management and comprehensive watershed governance.