Depth-domain tomography and Full Waveform Inversion (FWI) are the primary methods for subsurface velocity model building. Tomography relies on travel-time information, featuring high computational efficiency but with insufficient resolution. FWI utilizes the travel-time, amplitude and waveform information of seismic field data, providing high-precision, but is highly dependent on the initial model and seismic data quality, and prone to non-convergence issues during inversion. To comprehensively leverage the advantages of aforementioned techniques, we constructed a practical workflow and processing method for joint velocity modeling integrating depth-domain tomography and FWI under geological constraints. This approach involves using well logging data to constrain the construction of background velocity models and incorporating fault control information during both depth-domain tomography and full waveform inversion to constrain the updating direction. Comparative analysis of modeling results and actual data from the Dongpu exploration area demonstrates that the geologically constrained joint modeling method can: (1) obtain relatively accurate subsurface velocity models in complex fault-block areas; (2) suppress the influence of noise on inversion results; (3) achieve better fault positioning and seismic imaging in obscured zones; (4) significantly improve the focusing and flattening of image gathers; and (5) remarkably reduce well-to-seismic discrepancies.