The Heihe River Basin (HRB), located in the northwest of China, is a typical endorheic basin with unique and complex groundwater-land surface process. The dynamic interaction between groundwater and surface water is characterized by profound exchanges, illustrating the distinctive hydrologic features of the HRB. This study constructed a basin-scale, high-resolution (0.005 degrees and 1 h) integrated hydrologic model over the HRB (ParFlow-CLM-HRB), establishing the crucial connectivity for water and energy exchange across the atmospheric boundary layer, land surface, and subsurface layers. A comprehensive validation of multiple simulated hydrologic variables such as streamflow (SF), water table depth (WTD), evapotranspiration (ET) and land surface temperature (LST) was performed using data obtained from in-situ instruments or derived from spatial products. In general, the model showed strong performance in simulating SF at six hydrologic stations across the HRB, exhibiting a Spearman's correlation coefficient of rho > 0.6. Additionally, the model performed effectively for ET and LST, with biases of less than 0.5 and correlation coefficients exceeding rho > 0.7 when validated against eddy covariance and four-component radiometer observations. Although the model exhibited relatively pronounced discrepancies in simulating WTD compared to groundwater level well data, similar to 60 % of the simulated values fell within the observed range. Meanwhile, WTD simulations demonstrated comparable performance with two other referenced products across the alluvial fan and plain in the lower reaches; however, the results from ParFlow-CLM-HRB exhibited the greatest alignment with observed values, achieving the lowest root mean squared error (RMSE) of -1.08 m in the first sub-domain. ET and LST simulations also showed reasonable concordance with referenced remote sensing products; however, large discrepancies between simulations and remote sensing data were found in mountainous and desert regions. Finally, the model was inter-compared with existing models in the HRB, and the potential sources of model uncertainty were identified. Despite limitations in resolving deep aquifers and anthropogenic impacts, this work advances high-resolution modeling in data-scarce basins by prioritizing process-based validation and transparent uncertainty reporting, offering a benchmark for coupled groundwater-land surface process studies in complex terrains.