Background. Several anatomical studies provide evidence that green turtles (Chelonia mydas) possess the necessary anatomy for colour vision. Behavioural experiments have previously been conducted with newly emerged hatchlings, concluding that they are attracted to shorter wavelengths compared to longer wavelengths within a terrestrial environment, suggesting a possible attraction towards blue. This paper assessed the colour vision of hatchlings within an aquatic environment, and investigated whether the attraction for shorter wavelengths remains consistent within water, whether the colour saturation of the chromatic stimuli was an important factor, and whether rearing and testing individual animals in different coloured housing tanks has an impact on their visual choices. Methods. Forty-one hatchling green turtles were presented with a three-choice experiment where food was attached to three different coloured plates. The plates (blue, yellow, and red) were randomly arranged in the turtle’s tank and four different colour saturations were tested (100, 75, 50, and 25%). Turtles were individually placed into their housing tanks (coloured either red, white, blue or grey) with three different colour plates in front of them, from the same saturation level. The colour of the plate with food first approached and bitten by the turtle was recorded. Results. The colour of the tank in which an individual was reared, and where experiments were conducted, significantly influenced which food item was selected on the different coloured plates. While individual turtles preferred to select the food items associated with blue plates across the entire experiment (66.1% of the time compared to 18.2% and 15.7% for yellow and red plates respectively), the preference for blue plates was influenced by the colour of the rearing/experimental tank. Individuals raised in red, white or blue tanks appeared to consistently prefer food on blue plates, but there appeared to be no plate colour preference by turtles in grey tanks. There was no significant effect of either colour saturation or the spatial arrangement of the three colours within an experimental tank on colour choice, and no significant interaction between tank colour and colour saturation. Discussion. These findings confirm that the terrestrial preference towards shorter wavelength colours, such as blue, compared to longer wavelength colours remains consistent within an aquatic environment. This preference for blue continues even as the colour saturation reduces from 100% down to 25%, and the colours become darker. Thus, it is suggested that green turtle hatchlings have a strong attraction towards blue.