In climate science, ensemble modeling has emerged as a powerful tool for addressing the uncertainties inherent in individual climate models. This approach generates more robust and reliable predictions by harnessing the collective insights of multiple models. Nonetheless, the method of combining these models to derive an ensemble model remains an open question. To this end, the objectives of this research are twofold: (i) to introduce and evaluate the weighted average-correlation ensemble model for projecting minimum and maximum temperatures in Iran, and (ii) to assess near-term (2021–2040) trends across 95 synoptic stations using socio-economic scenarios derived from five models: GFDL-ESM4, MPI-ESM1-2-HR, IPSL-CM6A-LR, MRI-ESM2, and UKESM1-0-LL. The ensemble technique effectively reduces the Root Mean Square Error (RMSE) (1/3 − 1/10) associated with the individual models. The predicted values for the minimum temperature are more similar to the actual data than the maximum temperature. The results also indicate a significant increase in the minimum temperature compared to the maximum temperature during the base period. The distribution of the maximum temperature across the country is influenced mainly by its latitude. In contrast, the distribution of the minimum temperature is influenced by both the country’s major altitudes and latitudes. Surveys also indicate that, compared to the base period, there is an increasing trend in temperature for winter, spring, and autumn, while a decrease is observed during the summer. Notably, the increase in temperature is more pronounced during winter.