This study evaluates the efficiency of genomic prediction in multi-environment trials (MET) for single cross maize hybrids, focusing on grain yield and female flowering time. The research compares the performance of Genomic Best Linear Unbiased Prediction (GBLUP) with non-additive effects and machine learning methodologies (bagging, random forest, and boosting). The results show that both GBLUP and machine learning methods are effective in predicting hybrid performance, with GBLUP being particularly useful for optimizing variance component modeling. Machine learning methods, such as bagging, random forest, and boosting, can capture non-additive effects without prior assumptions, making them suitable for traits influenced by complex genetic architectures. The study highlights that predicting the performance of new hybrids not evaluated in field trials is more challenging than predicting hybrids in sparse test designs. Overall, the research demonstrates the potential of genomic prediction in maize breeding programs to reduce costs and increase the number of hybrid combinations evaluated.This study evaluates the efficiency of genomic prediction in multi-environment trials (MET) for single cross maize hybrids, focusing on grain yield and female flowering time. The research compares the performance of Genomic Best Linear Unbiased Prediction (GBLUP) with non-additive effects and machine learning methodologies (bagging, random forest, and boosting). The results show that both GBLUP and machine learning methods are effective in predicting hybrid performance, with GBLUP being particularly useful for optimizing variance component modeling. Machine learning methods, such as bagging, random forest, and boosting, can capture non-additive effects without prior assumptions, making them suitable for traits influenced by complex genetic architectures. The study highlights that predicting the performance of new hybrids not evaluated in field trials is more challenging than predicting hybrids in sparse test designs. Overall, the research demonstrates the potential of genomic prediction in maize breeding programs to reduce costs and increase the number of hybrid combinations evaluated.