<p>Mycotoxin contamination in wheat is strongly influenced by weather conditions, yet how contamination may evolve under future climate and socioeconomic change remains poorly understood at the European scale. Here, we develop a hybrid modelling framework combining machine learning and crop phenology to assess scenario-dependent changes in mycotoxin contamination. The framework integrates historical monitoring data with climate projections under multiple Shared Socioeconomic Pathways. Model evaluation shows strong performance for low-contamination conditions, while the ability to distinguish higher contamination levels remains limited due to class imbalance. Results indicate an overall increase in contamination risk under climate change, particularly for deoxynivalenol (DON), with relatively higher risk in coastal and northwestern Europe. These findings highlight the role of climate-driven shifts in crop phenology and weather conditions, and provide a scenario-based framework for exploring future mycotoxin risk patterns rather than precise quantitative predictions.</p>