We outline how redshift-space distortions (RSD) can be measured from the angular correlation function, w(θ), of galaxies selected from photometric surveys. The natural degeneracy between RSD and galaxy bias can be minimized by comparing results from bins with top-hat galaxy selection in redshift, and bins based on the radial position of galaxy pair centres. This comparison can also be used to test the accuracy of the photometric redshifts. The presence of RSD will be clearly detectable with the next generation of photometric redshift surveys. We show that the Dark Energy Survey (DES) will be able to measure f(z)σ8(z) to a 1σ accuracy of (17 ×b) per cent, using galaxies drawn from a single narrow redshift slice centred at z= 1. Here b is the linear bias, and f is the logarithmic rate of change of the linear growth rate with respect to the scalefactor. Extending to measurements of w(θ) for a series of bins of width 0.02(1 +z) over 0.5 < z < 1.4 will measure to a 1σ accuracy of 0.25, given the model f=Ωm(z), and assuming a linear bias model that evolves such that b= 0.5 +z (and fixing other cosmological parameters). The accuracy of our analytic predictions is confirmed using mock catalogues drawn from simulations conducted by the Marenostrum Institut de Ciències de l'Espai Simulations (MICE) collaboration.