More than 150 million people suffer from diabetes mellitus worldwide and this number is expected to rise substantially within the next decades. Despite its high prevalence, the pathogenesis of diabetes mellitus is not completely understood. Therefore, appropriate experimental models are essential tools to get more insight into the genetics and pathogenesis of the disease. Here, we describe the current efforts to establish novel diabetes models derived from unbiased phenotype-driven large-scale ENU mouse mutagenesis projects started a decade ago by using hyperglycemia as high-throughput screening parameter. Mouse lines were established according to their hyperglycemia phenotype over several generations thereby revealing a mutation as cause for the aberrant phenotype. Chromosomal assignment of the causative mutation and subsequent candidate gene analysis led to the detection of the mutations which resulted in novel alleles of genes already known to be involved in glucose homeostasis like glucokinase, insulin 2 and insulin receptor. Additional ENU-induced hyperglycemia lines are under genetic analysis. Improvements in the screen for diabetic animals are implemented to detect more subtle phenotypes. Moreover, diet challenge assays are being employed to uncover interactions between genetic and environmental factors in the pathogenesis of diabetes mellitus. The new mouse mutants recovered in phenotype-driven ENU mouse mutagenesis projects complement the available models generated by targeted mutagenesis of candidate genes, altogether providing the large resource of models required for a systematic dissection of the pathogenesis of diabetes mellitus.