Escalating climate change challenges in arid and semi-arid regions necessitate precise hydro-climatic evaluations for sustainable water resource management. This study aimed to simulate and project future temperature and precipitation changes across five synoptic stations in the Kashan Plain watershed (Natanz, Kashan, Meymeh, Kahak, and Delijan) using daily baseline data (2004–2024) and CMIP6 models. Outputs from eleven General Circulation Models (GCMs) were evaluated using five performance metrics (MAE, MSE, RMSE, MAPE, and R2), leading to the selection of four top-performing models: ACCESS-ESM1-5, CESM2, MPI-ESM1-2-LR, and BCC-CSM2-MR. Results revealed that CESM2 consistently exhibited the lowest error and highest coefficient of determination across all stations for temperature, serving as the benchmark model for temperature downscaling. For precipitation, the optimal model varied by station; however, MPI-ESM1-2-LR, BCC-CSM2-MR, and ACCESS-ESM1-5 delivered the best performance. Statistical downscaling was performed using LARS-WG8 under three scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5) for near-future (2021–2040) and mid-future (2041–2060) horizons. Results indicate that mean basin temperature in the mid-future will rise by 0.64°C (SSP1-2.6) to 1.29°C (SSP5-8.5), whereas warming signals in the near-future remain weaker (0.05–0.33°C). Contrary to common perception, mean basin precipitation shows a minor increase (1.5% to 7.2%) across most scenarios; however, this basin-wide average is driven by pronounced spatial heterogeneity. Natanz shows a significant increase (up to 25%), while Meymeh, despite its highest elevation, demonstrates a persistent decline (up to 15% under SSP5-8.5), whereas Kashan experiences minor changes (-4.4% to +7.7%). These findings indicate that a simple higher elevation equals greater precipitation relationship does not hold true in this basin, highlighting the necessity of station-specific monitoring and integrating these outputs with the WEAP water resource management model.