Gravity dock slope excavation is characterized by extensive excavation areas and prolonged construction periods.Traditional terminal displacement warning methods are difficult to achieve differentiated stability assessments across various stages of the long excavation period.As a static mechanical evaluation model,the limit equilibrium analysis method requires precise physical and mechanical property indices of the slope.However,when environmental or construction factors significantly influence the dock slope during excavation,the model becomes further distorted,exhibiting notable limitations.Based on this,this study establishes a multi-level dynamic early warning system for dock slopes.The system employs automated monitoring devices to conduct remote,automatic,and real-time monitoring of dock slopes,followed by similarity analysis of time-series data to identify abnormal environmental and construction factors during excavation.On this basis,the system invokes the PLAXIS interface for stability evaluation at each excavation stage,enabling multi-level dynamic warnings for gravity dock slopes.During the excavation of a dock slope in Zhoushan,the multi-level dynamic warning system accurately identified abnormal environmental factors,dynamically adjusted initial geotechnical physical-mechanical parameters,achieving advanced prediction of slope displacement and excavation stability at all levels,guiding construction personnel to protect in a timely manner,avoiding economic losses and casualties.This system provides new evaluation methods and advanced prediction directions for dock slope disaster prediction.