Understanding the relationship between weather extremes and crop productivity is the backbone of risk assessments on food security. The former is paramount in countries (i.e., Bhutan) where there are a limited number of impact assessments in agriculture. The analysis performed in this work highlights the trends in agricultural production under a changing climate and the attribution of yield change to a specific climatic hazard. We improve the understanding between the two (climate and yields) by applying an FAO eco-physiological model (PyAEZ) and by relying on state-of-the-art downscaled climate projections (CORDEX-CORE). We analyze climate change impacts on ten crops (maize, foxtail millet, buckwheat, wheat, rice, common beans, cabbage, white potatoes, carrots, and citrus) at national level. The main simulation findings point to a higher yield gain and loss under rainfed conditions as opposed to irrigated conditions by the end-century (2070-2099) compared to the baseline period (2010-2039), particularly under RCP 8.5 for white potatoes (+17.7%), wheat (+15.5%), cabbage (+12.8%), buckwheat (-6.6%), and citrus (-5.8%). The irrigation potential to modulate climate change impacts is reflected in a yield gain of +43.4% on average for all crops and RCP’s when compared to rainfed conditions. On average, weather extremes explain 28% and 33% of the yield variability over time, respectively under RCP’s 2.6 and 8.5. Overall, this study supports smallholder farmers, decision-makers, and project developers to implement adaptation solutions that minimize growing weather extremes such as heat-stress and dry-spells.