Abstract Optimizing wheat variety selection for high performance in different environmental conditions is critical for reliable food production and stable incomes for growers. We employ a statistical machine learning framework utilizing Gaussian process (GP) models to capture the effects of genetic and environmental factors on wheat yield and protein content. The GP approach is closely related to kernel linear mixed-effect and kernel regression models, which are commonly used for genotype × environment predictions. A notable advantage of the GP formulation is that it provides probabilistic predictions in closed form. Building on this, there is extensive literature for optimal decision-making and data acquisition strategies, opening the door to variety recommendation and experimental design strategies. By means of a wheat test case and using a novel dataset collected in Switzerland, we demonstrate that the GP approach delivers comparable or superior prediction results in comparison with state-of-the-art methods, while providing improved uncertainty quantification and supporting better informed decision-making.