Genomic prediction can improve the accuracy of estimated breeding values for traits driven by additive genetic effects within common settings but prediction of traits affected by non-additive genetic effects and GxE remains a challenge. Mechanistic growth models express growth performances in terms of nonlinear functional interactions between underlying latent traits and nutritional environmental effects. Assuming the latent traits are less affected by non-additive genetic effects and GxE, these models can capture certain non-additive genetic effects and GxE at the phenotype level and allow prediction at unobserved ages for longitudinal data, e.g. mature weight and mature feed intake. In this study, we developed a Bayesian hierarchical model to integrate a Gompertz model for body weight and feed intake into genomic prediction models for pigs. By predicting breeding values for biologically relevant underlying latent traits, these models have the potential to advance genetic improvement across populations and environments.