Genomic selection (GS) accelerates breeding by predicting breeding values using molecular markers. This review discusses GS and genomic-enabled prediction (GP) in plant breeding, focusing on their principles, models, and applications in cereal and legume crops. GS uses all molecular markers to predict breeding values, combining genetic and phenotypic data in a training population to estimate values in a testing population. GS reduces costs and time in breeding, as demonstrated in maize breeding, where it shortened the breeding cycle and increased grain yield. GP models, including genomic best linear unbiased prediction (GBLUP), have improved prediction accuracy for traits like grain yield, disease resistance, and agronomic traits. However, challenges remain, including the need for large marker datasets and the complexity of multi-trait and multi-environment interactions. Recent advances in machine learning and hyperspectral imaging have enhanced GP accuracy and enabled faster selection cycles. GS has shown significant genetic gains in maize, with cycles producing higher yields than conventional breeding. The integration of GS with high-throughput phenotyping and pedigree-based models improves prediction accuracy for traits like grain yield and disease resistance. Future prospects include expanding GS to legumes and using GS in germplasm enhancement programs to accelerate gene flow from gene banks to elite lines. Overall, GS offers a promising approach to improve crop productivity and adaptability in the face of climate change.Genomic selection (GS) accelerates breeding by predicting breeding values using molecular markers. This review discusses GS and genomic-enabled prediction (GP) in plant breeding, focusing on their principles, models, and applications in cereal and legume crops. GS uses all molecular markers to predict breeding values, combining genetic and phenotypic data in a training population to estimate values in a testing population. GS reduces costs and time in breeding, as demonstrated in maize breeding, where it shortened the breeding cycle and increased grain yield. GP models, including genomic best linear unbiased prediction (GBLUP), have improved prediction accuracy for traits like grain yield, disease resistance, and agronomic traits. However, challenges remain, including the need for large marker datasets and the complexity of multi-trait and multi-environment interactions. Recent advances in machine learning and hyperspectral imaging have enhanced GP accuracy and enabled faster selection cycles. GS has shown significant genetic gains in maize, with cycles producing higher yields than conventional breeding. The integration of GS with high-throughput phenotyping and pedigree-based models improves prediction accuracy for traits like grain yield and disease resistance. Future prospects include expanding GS to legumes and using GS in germplasm enhancement programs to accelerate gene flow from gene banks to elite lines. Overall, GS offers a promising approach to improve crop productivity and adaptability in the face of climate change.