The increasing frequency and intensity of climate extremes in recent decades highlight the need for reliable tools like Global Climate Models to project future changes in temperature, precipitation, and extreme climate indices at local and regional scales. Given the coarse spatial resolution of GCMs, downscaling approaches (statistical and dynamical) are widely applied to generate station-scale climate information. This study evaluates the performance of the latest version of the stochastic weather generator, LARS-WG8.0, in simulating temperature, precipitation, and climate extreme indices, and subsequently projects their future changes for the near-future period (2031–2050). To this end, daily observed minimum and maximum temperature and precipitation data from Mehrabad synoptic station, Tehran, were obtained from the IRIMO for the baseline period (1981–2010). The comparison of ETCCDI extreme indices revealed that LARS-WG8.0 adequately reproduces the central values and variability of the majority of extreme indices (approximately 12 out of 16 indices) for the study station. However, the most pronounced discrepancies were observed for indices associated with persistent and highly extreme events, including WSDI, CSDI, ID, and R99p. Future projections, derived from the outputs of five CMIP6 models under SSP1-2.6 and SSP5-8.5 scenarios, indicate that the number of summer days (SU) and tropical nights (TR20) will increase by 10–23 days and 12–30 days, respectively, relative to the baseline period. Moreover, maximum 1 day precipitation (RX1day) is projected to increase by up to 19% under the SSP5-8.5 scenario, while no significant changes are anticipated for the consecutive dry days index (CDD). These findings provide valuable insights for climate change adaptation planning and weather-related risk management in the Tehran metropolitan area.