1-Introdoction Occurrence of settlements associated with damaging adjacent structures is a dangerous phenomena due to tunnel excavation in the urban area. It may cause some dangers to the neighbors. Artificial intelligence as a new method applies to examine the relationship between different parameters and their impact on the geotechnical hazards incidents probability and severity. The amount of the settlement in EPB mechanized tunneling depends on several factors including 1) geological conditions and groundwater level, 2) tunnel characteristics (depth and diameter) and 3) drilling parameters (penetration rate, face pressure, back-fill grouting pressure, pitching angle of the machine). Due to the complexity of the settlement process in mechanized tunneling and the impossibility of using all of the useful parameters, the use of classical models such as regression and numerical methods is not efficient and has many problems (Suwansawat and Einstein, 2006). Therefore, the using of artificial intelligence models such as artificial neural networks and fuzzy methods frequently has been used to predict maximum settlement (Kim et al., 2001; Suwansawat and Einstein, 2006; Ocak and Seker, 2013; Show Fang et al., 2014; Moeinossadat et al., 2016; Nadiri et al., 2018). Artificial neural networks save time and decrease cost in modelling process. The analysis is based on artificial intelligence models by extracting the relationships between the factors affecting the settlement, such as tunnel depth and diameter, soil properties, and machine operation parameters. These models have been used to predict ground surface settlement in a number of tunneling projects. (Inanlou and Ahanghari, 2010; Rezazadeh Anbarani et al., 2013; Jafari et al., 2013; Santos and Celestino, 2008; Ocak and Seker, 2013; Dindarloo and Siami Irdemoosa, 2015; Mohammadi et al., 2015; Camos et al., 2016; Bouayad and Emeriault, 2017). In this research, in addition to solving previous research problems, a model is proposed that its results can be generalized to the other projects all over the world. The study area was selected as part of the Tabriz metro line 1, between the Qunqa and Gazran stations, which has suitable conditions in viewpoints of geological characteristics and tunnel specification that is very common and similar to other projects in this field. A fuzzy modeling method has been used to predict a maximum settlement in the study area, and its results are compared with the results of artificial neural networks which is used in previous studies. Also, considering the ability of each model, the combination of these models can take advantage of the simultaneous benefits of both models (Nadiri et al., 2013, 2014, 2015, 2018). So in this research, the Neuro-fuzzy model is used to combine individual artificial intelligence models to use the benefits of both models simultaneously...