Methane production is a trait of increasing importance for the Australia sheep industry, it is being recorded on industry and reference flocks, with project breeding values provided to participating breeders. The current sheep index development in Australia uses traits related to production and reproduction, and from this methane production can be predicted and included in a sustainability index if desired. The purpose of this study was to estimate genetic and phenotypic correlations between methane production and commonly recorded production traits, and to assess the benefit of including direct measurements of methane using genomic prediction in a selection index. This analysis included 6,086 purebred Australian Merino sheep with methane production recorded using portable accumulation chambers. The estimated heritability for methane production was 0.19 ± 0.03. A series of bivariate animal models were used to estimate covariances for methane and commonly recorded production traits. The genetic correlations between methane and the following production traits were: weight at the time of recording (0.36 ± 0.08), weaning weight (0.47 ± 0.05), adult weight (0.34 ± 0.08), yearling c-site fat depth (-0.09 ± 0.08), yearling greasy fleece weight (0.21 ± 0.07), yearling fibre diameter (-0.02 ± 0.07), and yearling faecal egg count (-0.05 ± 0.10). The genetic correlations between methane and the production traits were within the parameter space of published literature. The estimated genetic parameters were used along with selection index theory, to predict selection accuracy and response to selection. Each of the production traits was assumed to have a genomic prediction accuracy similar to having 10 recorded progeny. The assumed genomic prediction accuracy for methane production was 0.40 (based on heritability). Methane production had 100% of the selection pressure for both with and without genomic prediction. The EBV accuracy for methane predicted only based on correlated traits (0.38) was considerably lower compared to adding a genomic prediction for methane (0.52). These results indicate there is a benefit to developing a reference population for methane rather than relying on correlated traits, however, this is dependent on an assumed genomic prediction accuracy of 0.40. In this study, only commonly recorded traits were analysed, and only single trait selection was investigated. A balanced selection index that contains relevant economic values for multiple traits will be more appropriate. Further research is expanding the analysis to include other breeds and a broader range of traits, including reproduction traits, and to integrate these findings into index development for Australian sheep. These results underscore the value of continued methane measurement strategies and the need for increasing the reference population. The results also demonstrate that direct selection can mitigate methane production more effectively than indirect selection based purely on correlated responses.