Abstract: In order to predict the shield tunnel surface settlement more accurately and effectively,this paper proposes an empirical modal decomposition (EMD),sparrow search algorithm (SSA) and extreme learning machine (ELM) combination of shield tunnel surface settlement prediction method.Firstly,EMD is used to decompose the subsidence sequence into trend and fluctuation vectors to fully extract the effective information of the sequence;secondly,the sparrow search algorithm is improved by introducing Cubic chaotic mapping to initialize the population and adaptive factor to optimize the searcher’s position formula,and a chaotic adaptive sparrow search algorithm (CASSA) is proposed;finally,CASSA is used to The weights and thresholds in the extreme learning machine are optimally searched to construct a CASSA-ELM prediction model.The fluctuation components and trend components decomposed by EMD are predicted one by one,and the prediction results are superimposed and reconstructed to obtain the predicted final settlement.Taking a shield interval of Nantong Railway Line 1 as an example,the combined EMD-CASSA-ELM model improves the prediction accuracy by 3.86% compared with the traditional ELM model.It is found that the combined EMD-CASSA-ELM model can effectively improve the prediction accuracy of surface settlement,verify its good generalization ability,and provide a new means for safety monitoring.