Owing to its improved accuracy, genomic selection, which utilizes genetic markers across the whole genome to predict the genetic merits of each individual, has become a key technique in pig and other livestock breeding. This chapter reviews developments in optimizing different models used in genomic selection. It reviews improving classical algorithms using techniques such as Genomic Best Linear Unbiased Prediction (GBLUP). It also discusses improving BLUP model approaches e.g. via the single-step GBLUP (SSGBLUP) model and differing ways of weighting genetic markers, as well as appropriate software for particular modelling approaches. The chapter also assesses ways of improving the predictive accuracy of Bayesian, e.g. by using differing assumptions on distribution of marker effects and ways of accounting for unknown parameters, in order to optimise the potential of genomic selection.