Context.Gamma-ray binaries are thought to be composed of a young pulsar in orbit around a massive O or Be star with their gamma-ray emission powered by pulsar spin-down. The number of such systems in our Galaxy is not known.Aims.We aim to estimate the total number of gamma-ray binaries in our Galaxy and to evaluate the prospects for new detections in the GeV and TeV energy range, taking into account that their gamma-ray emission is modulated on the orbital period.Methods.We modelled the population of gamma-ray binaries and evaluated the fraction of detected systems in surveys with theFermi-LAT (GeV), H.E.S.S., HAWC and CTA (TeV) using observation-based and synthetic template light curves.Results.The detected fraction depends more on the orbit-average flux than on the light-curve shape. Our best estimate for the number of gamma-ray binaries is 101\hbox{$_{-52}^{+89}$} systems. A handful of discoveries are expected by pursuing theFermi-LAT survey. Discoveries in TeV surveys are less likely. However, this depends on the relative amounts of power emitted in GeV and TeV domains. There could be as many as ≈ 200 HESS J0632+057-like systems with a high ratio of TeV to GeV emission compared to other gamma-ray binaries. Statistics allow for as many as three discoveries in five years of HAWC observations and five discoveries in the first two years of the CTA Galactic Plane survey.Conclusions.We favour continuedFermi-LAT observations over ground-based TeV surveys to find new gamma-ray binaries. Gamma-ray observations are most sensitive to short orbital period systems with a high spin-down pulsar power. Radio pulsar surveys (SKA) are likely to be more efficient in detecting long orbital period systems, providing a complementary probe into the gamma-ray binary population.