Mountain climate assessments in Central Asia are sensitive to the choice of gridded climate dataset, particularly in complex terrain where meteorological observations are sparse. This study compared annual air temperature (1980–2019) and precipitation (1980–2018) series extracted from CHELSA v2.1 and WorldClim 2.1 historical monthly data at three mountain sites of southeastern Kazakhstan and the Tien Shan: Kegen Pass, Tekeli Gorge and the Terskey Alatau–Saryzhaz site. Trends were assessed with the Mann–Kendall test with trend-free pre-whitening and Sen's slope, and the temperature series were extended to 2040 with ARIMA models. Both products indicated warming at all three sites, but its magnitude differed. WorldClim showed nearly uniform trends of 0.31–0.38 °C per decade (p < 0.001), with decadal means rising by 0.9–1.0 °C between 1980–1989 and 2010–2019, whereas CHELSA showed weaker and more variable trends of 0.14–0.23 °C per decade (0.3–0.7 °C between the same decades); at Kegen Pass the CHELSA trend was only marginally significant. WorldClim was 1.2–4.2 °C warmer than CHELSA in absolute terms. Divergence was greater for precipitation: at Tekeli Gorge CHELSA gave about twice the WorldClim amount (mean 94 vs 48 mm per month), and a significant increase was found only in CHELSA at the Terskey Alatau–Saryzhaz site (+8.6 mm per month per decade, p < 0.001) and, more weakly, at Kegen Pass. ARIMA forecasts for 2040 exceeded the 2010–2019 mean by 1.0–1.1 °C in the WorldClim series and by −0.1 to +0.6 °C in the CHELSA series, but the 95% prediction intervals (about ±1 °C) overlapped, so the forecast differences are not statistically distinguishable. The direction of long-term warming is therefore robust to product choice, whereas absolute values, warming magnitude and precipitation remain product-dependent and should be treated as structural uncertainty in mountain-climate and environmental assessments.