The expansion of the underground tunnel network is calling for new methods to allow for their monitoring in an efficient way. A new approach to assess the safety level of segmental lining of tunnels excavated in rock based on discrete monitoring points is developed. Artificial Neural Networks (ANNs) are deployed for the prediction of quantities of interest, chosen for the assessment of the utilization level of the lining, based on input strains. A finite element (FE) model of the lining is created to generate the data required for the training of the ANNs. Eventually, two benchmark scenarios are defined in order to test the approach, by creating FE models representing the excavation process and the lining installation. The predictive model is fed with the input quantities of the testing scenarios and a comparison among predictions and reference quantities is drawn.