Erdos distinct distances problem asks for the lower bound of number of distinct distances between pairs of points from any finite point sets of given size. The problem in the Euclidean plane was revolved by L. Guth and N. H. Katz in 2011. We study the problem in hyperbolic surfaces. The key in our work is to introduce an invariant (we call it "geodesic cover") for Fuchsian groups, which summons copies of fundamental polygons in the hyperbolic plane to cover pairs of representatives realizing distances in the corresponding hyperbolic surface. Then we use estimates of this invariant to study the distinct distances problem in hyperbolic surfaces. Especially, for S from a large class of hyperbolic surfaces, we establish the nearly optimal bound >= C_S N/\log N for distinct distances determined by any N points in S, where C_S>0 is some constant depending only on S. In particular, for S being modular surface or standard regular of genus >=2, we evaluate C_S explicitly.