Abstract: Tunnel deformation monitoring plays an important role in ensuring tunnel safety. For this study, collected deformation data were analyzed and predictions were established using the grey theory GM model; it was found that grey theory has certain predictive effects on tunnel deformations. It was also discovered that there are some differences in the prediction accuracy of different sample data. Therefore, the quadric parameter fitting method was proposed to improve these differences, and the results indicate that this method improves low-accuracy prediction models, but is not so significant for high-accuracy prediction models.