Abstract: To ensure balanced control of earth pressure during the shield machine tunneling process, this study proposes an intelligent optimization control strategy based on shield machine tunneling parameters. The strategy is based on an advanced predictive model framework, aiming to construct an optimized control model by minimizing the deviation between the set earth pressure value and the predicted value. The optimization control process is divided into three steps: first, discrete wavelet transform is used for denoising the raw construction data; then, a one-dimensional convolutional neural network and a long short-term memory neural network are utilized for feature extraction and multi-step prediction of earth pressure; finally, the evolutionary particle swarm optimization algorithm is employed for online optimization of tunneling parameters such as propulsion speed and screw conveyor speed, thereby achieving balanced control of earth pressure throughout the tunneling process. Experimental results show that this strategy can accurately track the set earth pressure and control earth pressure both inside and outside the sealed chamber, and significantly enhance the safety and efficiency of shield machine tunneling operations.