Home range analysis is a powerful tool for identifying priority areas for conservation, but estimating the home range for many species is still challenging. In particular, highly mobile species may use different areas at different times (e.g. summer or winter), so temporally biased location data may only partially represent their home range. We investigated the temporal patterns in habitat use of green turtles Chelonia mydas (n = 52) and loggerhead turtles Caretta caretta (n = 20) at longer (>1 yr) and shorter (<1 yr) scales. The study was conducted in subtropical and tropical foraging habitats along the Queensland coast of Australia between 1991 and 2015. Each turtle was tracked by a satellite-linked tag for the effective life of the device; 3 turtles were tracked twice. Mark-recapture studies were also conducted intermittently. Single satellite-tag deployments confirmed site fidelity to a foraging habitat for up to 2.5 yr in green turtles and 2.7 yr in loggerhead turtles. Further, combining satellite telemetry and mark-recapture records indicated much longer periods of foraging residency, up to 17 yr for green turtles and 23 yr for loggerhead turtles. No tracked turtles made substantial changes in their foraging range between years. Within the long-term home range, subtropical turtles tended to shift their foraging areas seasonally. Consequently, for many turtles, the existing conservation legislation provided protection in some seasons but not others. Our results emphasise the importance of protecting areas according to the turtles' use of space, with careful consideration given to identify temporal trends in their habitat selection.