This paper deals with the finite-time leader-following consensus problem for a class of multi-agent systems with nonlinear dynamics and undirected communication topology. A distributed adaptive nonlinear control law is proposed based on the relative state information between neighboring agents, which solves the finite-time leader-following consensus problem. Compared with some existing results in the literature, we do not put any any intrinsic restrictions (such as Lipschitz or Lipschitz-like conditions) on nonlinear functions and the adaptive consensus protocol is in a distributed fashion. A numerical example is given to verify the theoretical analysis.