Breeding for high baking quality in wheat is crucial in Switzerland, and a lot of effort is required to meet the high expectations of stakeholders and consumers. Assessment of baking quality is time-consuming and expensive. Accordingly, efficient and precise predictions of baking quality-related traits are essential for breeding. The goal of this project is to improve efficiency of baking quality estimation of wheat breeding lines by emphasizing increasing prediction accuracy. Our current dataset of phenotypic and genotypic data available on a set of about 1800 breeding lines will be combined with new data to study genetic factors underlying baking quality. Genomic selection models and machine learning models will be developed to predict baking quality traits. We place particular emphasis on breeding varieties with high baking quality under reduced nitrogen fertilization as well as the development of genetic markers for high molecular weight glutenin alleles. As a final step, we aim to develop an improved process to predict baking quality in early generations of the breeding cycle and implement it in the wheat breeding programs.