<p>Climate change represents a growing challenge to Turkey’s climatic stability, underscoring the necessity for reliable forecasting methodologies to facilitate adaptation strategies and informed policy development. The present study examines long-term trends in national-scale temperature and precipitation series between 1902 and 2022 and produces future projections for the period 2022–2040 using an integrated structure of statistical and artificial intelligence-based models. Four distinct models were utilized in the prediction process: Vector Autoregression (VAR), Extreme Learning Machine (ELM), Multilayer Perceptron (MLP), and Nonlinear Autoregressive Exogenous (NARX). These models were then compared with CMIP6-based Shared Socioeconomic Pathways (SSP1-1.9, SSP2-4.5, SSP5-8.5) scenarios. The Mann–Kendall and Pettitt tests revealed a statistically significant warming trend after 1980, as well as structural breaks in temperature in 1994 and in precipitation in 1935. The findings of quantitative assessments (R², RMSE, NSE, KGE) demonstrate that ELM and MLP models exhibit superior accuracy (R² ≥ 0.95, RMSE ≤ 0.25&#xa0;°C) in temperature predictions. Projections indicate that the average annual temperature in Turkey will reach approximately 13.2–13.6&#xa0;°C by 2040, and this trend shows a high degree of consistency with the medium emission scenario SSP2-4.5. Conversely, the performance of models in precipitation forecasts was found to be less consistent, owing to elevated temporal and spatial variability. Statistical tests (ANOVA and Kruskal–Wallis) demonstrate that there are significant differences between temperature prediction models, but no clear superiority in precipitation predictions. In conclusion, this study presents one of the first national climate projection frameworks for Turkey on a century scale, combining data-driven modelling with scenario-based projections to provide quantitative contributions to climate adaptation policies and regional planning processes.</p>