Abstract: In order to reduce the occurrence of disasters such as rock fall and landslide during the construction of road tunnels,this paper establishes a road tunnel instability risk assessment model based on XGBoost algorithm on the basis of giving full play to the value of engineering data.Six indicators,namely,weathering degree of rock body,surrounding rock level,integrity of rock body,structural surface bonding degree,groundwater outflow and excavation method,are selected as the assessment indicators of the instability risk assessment model of the road tunnel,and the risk assessment is achieved by quantitatively assigning the values to each qualitative indicator.Confusion matrix and accuracy rate are introduced to test the training effect of the test set and prediction set of the instability risk assessment model.The risk assessment model is validated by comparing the results of the Dahuashan Tunnel with the results of the actual project,and the validation results show a high degree of consistency,which successfully warns the occurrence of surrounding rock instability disaster.Based on Visual Basic,the “Road Tunnel Instability Risk Assessment System (RCIAS)” was developed independently,which realises accurate and fast instability risk assessment in the construction process of road tunnels.