The simulation of groundwater flow is widely used to understand dynamic groundwater processes and frequently serves as the basis for quantitative analysis and management of groundwater resources. Reliable groundwater flow simulations require a sound understanding of the hydrogeological structure of the subsurface materials. However, a detailed model of the hydrological structure of thick unconsolidated basin sediments is often limited by their inaccessible nature, especially in the deep domain. As a result, deep regional flow is generally neglected in traditional groundwater flow simulations. This study presents an integrated approach for developing a model of hydrogeological structures using multiple data sources, including audio-magnetotelluric data (AMT), borehole data, and the interpreted lower boundary of Quaternary sediments. The approach overcomes the disadvantages of using the depth of boreholes as the lower boundary of the hydrogeological system and may be used to analyse the effects of hydrogeological structures on groundwater flow simulations. The model, which utilized sequential Gaussian simulation (SGS) and natural neighbor interpolation (NNI), was evaluated by application to a typical cross-section through the Yinchuan Basin, China, to a depth of 1,700 m. The horizontal hydraulic conductivity (K_x) field was calibrated using observation wells and the PEST pilot points (PPP) method. The effects of the developed hydrogeological structure on groundwater flow simulations were evaluated for the entire thickness of Quaternary sediments (complete-domain model, CDM) as well as for sediments above the maximum depth of the boreholes (shallow-domain model, SDM). In addition, the sensitivity of the K_x field on groundwater flow patterns was analysed. The results show that the proposed approach can characterize the spatial distribution of lithologic units and the heterogeneity of hydrogeological parameters in the deep domain (>300 m), as well as avoid the multi-solution problem of inferring hydrogeological parameters from geophysical data. It can also improve the accuracy of the hydrogeological structure model and the reliability of groundwater flow simulations. The adjustment of the hydrogeological structure affects the flow budget as well as the pattern of the groundwater flow systems (including the type of system, and their number, location, and extent). Changes in the K_x field affect the local flow field, whereas adjustment of the hydrogeological structure model affects the global flow field.