In order to solve the problem of realizing the autonomous deviation correction of the shield machine, based on the actual engineering data, a shield correction control method based on random forest algorithm (RF) and genetic algorithm (GA) was proposed. The prediction model is combined with the optimization model, and the target deviation value is input to invert the output shield correction parameter value, which further improves the automation of shield correction, and then compares it with the actual data to verify the feasibility of the model. The results show that the R2=0.876 and MSE=6.448 of the RF algorithm prediction model, and MAE=2.04, the accuracy control can meet the requirements of the construction specification. The RF-GA shield tunneling correction model can control the deviation below 7mm in real-time, and the average accuracy of the output shield tunneling parameters and actual data is above 90%. In the comparative experiment, the variation pattern of shield tunnel parameters output by the control model is consistent with the actual pattern. It provides a new idea to realize the attitude control of shield machine and study the variation law of shield machine parameters in practical engineering.