<p>Wellbore instability continues to pose a significant challenge during drilling operations, particularly under intense stress and complex geological structures that result in a narrow MWW (mud weight window). Current techniques utilizing finite-element simulations for geomechanical modeling only can produce 1D (one-dimensional) or 2D (two-dimensional) MWWs due to high computational costs and well sparsity in the presence of complex geology. To address these issues, this study adopts an interdisciplinary approach that combines geophysics and rock mechanics. A geophysical-driven workflow is built for estimating the large in situ stress models along with 3D (three-dimensional) MWWs for wellbore stability analysis. The process begins with seismic inversion, integrating pre-stack seismic data with rock property logs to characterize the 3D elastic parameters of rock formations from the surface down to the target depth. Subsequently, curvature attributes are incorporated into the traditional MEM (mechanical earth model) framework to construct a large high-precision 3D in situ stress model. Finally, by applying the Mohr–Coulomb criterion and tensile failure criterion, these 3D pressure models (for pore pressure, collapse pressure, breakdown pressure, and loss pressure) are then computed to determine the MWW. A large in situ stress model covering 600km² with a total of 1.36 billion grids is established in the Tarim Basin, China, achieving prediction of the safe MWW from 1D logging data to the 3D working area in the study target. The blind well tests demonstrate the reliability of our MEMs, with an accuracy rate of 93.79%, and the resulting MWWs show a strong correlation with drilling complexity events. This evidence validates the proposed workflows as potential supportive tools for drilling operations.</p>