The objective of this study is to evaluate the impact of critical variables such as excavation depth (Depth), vertical displacement (Displacement Uz), pore water pressures, and suction efficiency on the effective stress σ′zz during the excavation process. To achieve this goal, the Finite Element Analysis (FEA) method was employed to simulate changes in geotechnical factors under real-world conditions. Simultaneously, Artificial Neural Networks (ANN) were applied to process and analyze data from FEA, aiming to improve the accuracy of effective stress prediction and optimize computational processes. The results from the study indicate that excavation depth and pore water pressures have the most significant impact on σ′zz, particularly at greater depths where the increase in displacement and pore water pressures leads to a substantial reduction in effective stress. The ANN model demonstrated its capability to accurately predict σ′zz with an R2 value of 0.974, while also clearly identifying the importance of each input variable. The combination of FEA and ANN not only enhanced accuracy but also reduced processing time, contributing significantly to the optimization of excavation design and execution. This study holds great significance for the field of construction and soil mechanics, as it provides a more accurate simulation and prediction solution for soil stresses, especially in deep excavation projects.