Climate change, soil degradation, and rising fertilizer costs threaten global food security. Traditional maize varieties, known as landraces, offer valuable genetic diversity for addressing these challenges, potentially carrying beneficial alleles for stress tolerance due to their adaptation to diverse agro-climatic conditions. However, they remain underutilized in modern breeding programs due to poor characterization, genetic heterogeneity, and lower agronomic performance. The MineLandDiv project aims to bridge this gap by integrating genomic data, high-throughput phenotyping, and multi-environmental field trials of 300 temperate landraces to better understand the resilience of landraces to abiotic (heat, drought, cold, nitrogen) and biotic (Corn borer) stresses. This comprehensive characterization will (i) identify stress-tolerant alleles through genome-wide association studies, thereby expanding the genetic diversity of modern breeding programs, and (ii) predict the agronomic performance of a broader range of landraces across various environments by applying genomic selection and genomic offset models based on phenotypic, genomic, and environmental data.