Abstract: In various prediction methods for ground surface settlement caused by tunnel excavation, it is assumed that the convergence mode of the tunnel is symmetric, ignoring the influence of asymmetric convergence of the tunnel. To predict ground surface settlement caused by biased tunnels, a biased convergence mode for tunnels is proposed for the first time with defining corresponding biased parameters. Based on stochastic medium theory, using coordinate transformation and integral numerical processing, a prediction model for ground surface settlement caused by biased tunnels is obtained. The applicability of this method is verified through practical projects, and the influence of relevant parameters on ground surface settlement is analyzed. Compared to the traditional convergence model, this convergence model obtains the maximum settlement and the location of the maximum settlement point more accurately, as well as the overall trend of ground surface settlement.