During the excavation process of a tunnel shield machine， a single attitude adjustment parameter remains unchanged， resulting in a significant deviation in the construction axis of the machine. Therefore， a method for controlling the excavation posture of tunnel shield machines based on autonomous correction algorithms is proposed. Using the triangular grid algorithm， the spatial environment map of tunnel construction is segmented and a grid map of excavation space is established. Analysis of the linear displacement process of each shield tunneling single ring excavation in a grid map is made to obtain the specific posture of the shield tunneling machine during construction. Compared with the expected construction and excavation posture， the current shield machine construction posture deviation is calculated. Combining PID control algorithm and BP neural network， a new autonomous correction algorithm is designed. Based on real-time posture deviation， the posture adjustment parameters are adaptively adjusted to achieve shield machine construction posture correction control. The test results of the example show that after the application of the new research control method， the axis deviation of shield tunneling machine construction does not exceed 10 mm， which meets the requirements of excavation posture control standards.