Mountain vegetable systems are critical for income generation and nutritional security in the Indian Himalayas, but accelerating climatic warming is increasingly threatening their long-term sustainability. To evaluate climate–production dynamics in Himachal Pradesh during 1996–2025, this study employed a multi-method framework integrating Mann–Kendall (MK) trend analysis, Sens’ slope estimation (SSE), compound annual growth rate modelling (CAGR) modelling, decomposition analysis (DA) , spearman correlation (SC) and extreme gradient boosting (XGBoost)-based machine learning. Results revealed substantial warming, with maximum and minimum temperatures increasing by approximately 0.12 °C year-1, accompanied by a sharp decline in chilling hours (-31.88 hr year-1). Although vegetable production increased considerably during the study period, decomposition analysis showed that production growth was driven primarily by rapid expansion of cultivated area, with tomato and pea production increases attributable to area effects by 96.23 % and 81.43 %, respectively. In contrast, productivity growth remained weak or negative for several climate-sensitive vegetables. Machine-learning analysis identified critical thermal thresholds above 22–23 °C, while exploratory (XGBoost) models demonstrated satisfactory performance, with R² values ranging between 0.78–0.89, mean absolute error (MAE) values between 5–8 % and mean absolute percentage error (MAPE) values below 10 %. Overall, the findings suggest that current vegetable production growth in Himachal Pradesh reflects structural buffering through area expansion rather than genuine productivity-based climate resilience, highlighting the need for climate-adaptive and productivity-centred management strategies in Himalayan agroecosystems.