It has been proposed that during the formation of the first generation stars there might be a dark star phase in which the power of the star comes from dark matter annihilation. The adiabatic contraction process to form the dark star would result in
a highly concentrated density profile of the host halo at the same time, which may give enhanced indirect detection signals of dark matter. In this work we investigate the extragalactic $gamma$-ray background from dark matter annihilation with such a dark star formation scenario, and employ the isotropic $gamma$-ray data from Fermi-LAT to constrain the model parameters of dark matter. The results suffer from large uncertainties of both the formation rate of the first generation stars and the subsequent evolution effects of the host halos of the dark stars. We find, in the most optimistic case for $gamma$-ray production via dark matter annihilation, the expected extragalactic $gamma$-ray flux will be enhanced by 1-2 orders of magnitude. In such a case, the annihilation cross section of the supersymmetric dark matter can be constrained to the thermal production level, and the leptonic dark matter model which is proposed to explain the positron/electron excesses can be well excluded. Conversely, if the positron/electron excesses are of a dark matter annihilation origin, then the early Universe environment is such that no dark star can form.
