Abstract: In order to realize attitude control of the shield tunnelling machine and ensure safety and quality of tunnel construction, in this study, a hybrid intelligent framework incorporating random forest (RF) and non-dominant sorting genetic algorithm-Ⅲ(NSGA-Ⅲ) is proposed. Based on the Wuhan Metro project, 17 influencing factors are selected as input variables, and the nonlinear mapping function relationship between input parameters and shield attitude is established by the RF algorithm, which is used as the fitness function of NSGA-Ⅲ, and then the key influencing factors are determined by the RF algorithm. A multi-objective intelligent optimization model of RF-NSGA-Ⅲ is established to minimize the absolute value of shield attitude parameter. Based on the proposed optimization principle, case study is conducted to test the applicability and effectiveness of the proposed method. The results show that the prediction model obtained by training and simulating measured engineering data using the RF algorithm has high accuracy. With the developed RF-NSGA-Ⅲ intelligent algorithm, a remarkable optimization and control effect of shield attitude is obtained.