queezing is one of the most important hazards in the tunneling projects that can lead to tunnel instability, increasing costs and stopping of the tunneling operation. This phenomenon often occurs in weak rock mass and high depth. Accurate prediction of squeezing has play an important role on selection of tunneling method and design of support system in design stages of tunneling. During past years, many methods have been developed to predict the squeezing phenomenon, most of which can be divided into experimental and semi-experimental methods. In recent years application of data mining methods for prediction of squeezing phenomenon has been increased remarkably. Survey have shown that the use of data mining techniques results in development of high-performance models. Therefore, in this study, using data mining based classification methods, models have been presented for squeezing prediction. The main focus of this thesis is on C5.0 decision tree and k-nearest neighbor (kNN). After developing the models, using mentioned methods, their performance were compared with other data mining methods that previously have been employed by other researchers, such as logistic regression and support vector machine. The results show that the developed models using C5.0 and k-nearest neighbor methods have higher performances in comparison to the common experimental equations and other data mining methods. Therefore the output of these methods can be applied for squeezing prediction in tunneling projects with an acceptable accuracy