Calculating the genetic merit of animals using genomic Estimated Breeding Values (gEBVs) has rapidly accelerated many terrestrial animal breeding programs. However, its adoption in aquaculture can be prohibitive due to the large amounts of resources (i.e., cost and time) required to genotype and phenotype large numbers of animals that are commonly reared together in each production cycle. Pooled genotyping is a strategy that can help circumnavigate this barrier, by reducing the resources required, while maintaining high levels of gEBV prediction accuracy. The purpose of this study is to test accuracies of gEBVs obtained from pooled records compared to individual records. The outcomes of this research will be of importance particularly for mass spawning and communal reared aquaculture species that require the ability to estimate gEBVs accurately and cost effectively to achieve industry advancement. This study used 5,273 black tiger prawn individuals (Penaeus monodon) genotyped and phenotyped for body weight across eight commercial grow-out ponds with overlapping families. The pooled gEBVs were calculated under two scenarios. Firstly, gEBVs were calculated based on individual progeny records and then pooled gEBVs estimated based on i) ranked phenotype and ii) full sib & half sib relationships; pool sizes consisted of 2, 5, 10, 15, 20 and 25 individuals per pool. The gEBVs were statistically derived from a single step genomic best linear unbiased prediction model (ssGBLUP). The parent gEBV accuracy pool scenario was measured by Pearson correlation coefficient between parent gEBV from pooled progeny and parent gEBV from individually genotyped progeny. The results indicated that increased pool size decreased the gEBVs accuracy of the unknown parent phenotypes. Higher correlation was achieved in pool sizes below 10 for both pool scenarios, whereas correlation further declined as pool size increased. However, the correlation was slightly lower in full sib & half sib pools. Pool sizes less than 10 produced gEBVs and predicted phenotypes that are more similar to those calculated on individual data (Table 1). Overall, this study provides the first results of genomic prediction on farm in P. monodon suggesting that using an optimum range pool of 10 to 15 individuals can reduce genotyping expenses (~10-fold reduction) whilst maintaining accurate estimation (~85% of individual gEBV correlation). These results could find potential applications for use in genomic selection in commercial breeding programs.