Abstract: Fast classification of surrounding rock mass during construction period is the key measure for construction security and engineering quality. This paper combines the surrounding rock mass classification work during tunnel group construction in Jihuang and Ningji expressway. It proposes a new classification system as a kind of improved national standard BQ classification method. The system is based on many field tests and indoor experiments. The quick field testing method of each classification index is presented. Furthermore, the new classification system is used to cope with the surrounding rock classification in-situ. The classification results are served as the training samples of the evolutionary support vector regression(SVR)algorithm in order to establish the intelligent classification model. Finally, the intelligent model is transferred to an elementary function mathematical model according to SVR theory for the purpose of convenience application. The accuracy of this mathematical classification model is verified by classification examples of tunnel group. Thus, the system offers a new simple fast classification method for surrounding rock mass during tunnel construction period.