Abstract: Considering on the monitoring problem of foundation pit settlement, four prediction models were proposed. Model 1 was a traditional grey Verhulst model. Model 2 was an optimized grey Verhulst model which used the reciprocal transformation and the optimization model without constraint to acquire the parameters of the background value and initial value. Model 3 was an optimized discrete grey Verhulst model that was built on the reciprocal transformation of the original data sequence by discretization method. Model 4 was an optimized discrete grey Verhulst model constructed through the metabolic method. The proposed models were used to predict the settlement of the foundation pit of a building in Longcheng Industrial Park in Shenzhen. The comparison between the calculated results and the measured results shows that the optimized gray discrete Verhulst model with the metabolic method has higher prediction accuracy and stability, and could better describe the settlement.