A complete diallel cross involving four strains of red tilapia (Oreochromis spp.) from Malaysia, Stirling, Taiwan and Thailand was performed with the aim to establish a foundation population for genetic improvement. The mating involved 16 parental breeders per sex per strain, producing 64 full-sib families in total. Statistical analyses were carried out on data consisting of records from 1280 individuals reared in both freshwater (0 ppt) and saline water (30 ppt) environments. Among the purebreds, and across the testing environments, the Malaysian strain exhibited the highest additive genetic performance for harvest weight. Whereas the Stirling strain was the poorest (7.4% above and 13.4% below the overall mean of pure strains, respectively). The average level of heterosis was generally low (4.2%) and the average of all crossbreds was not statistically different from the mean of the pure strains. Overall, the growth performance of red tilapia was significantly lower in saline than in freshwater environments (Pb0.001). The high correlation between additive genetic and total performance suggests that improvement of red tilapia can be effectively based on the exploitation of additive genetic variation (i.e. through selective breeding). A synthetic base population was therefore formed with the best performing individuals regardless of their genetic makeup and of the environments tested. Strategies for the future breeding program in red tilapia are discussed.