A distributed fast finite-time bipartite tracking consensus problem for first-order nonlinear multi-agent systems with uncertain disturbances is studied, where the leader has an external input unknown to all followers. Based on global finite-time stability theory, algebraic graph theory and variable transformation, a static distributed non-smooth protocol is proposed and analyzed. It is proved that in a nonlinear multi-agent system with disturbance, all the followers can track the time-varying expected state rapidly in a finite time. The limitation of this protocol is that the control gains depend on some global information, such as the spectrum of the Laplacian matrix. In order to eliminate this limitation, an adaptive distributed protocol is further designed. Theoretical analysis shows that the multi-agent system can achieve fast finite-time bipartite tracking consensus on the premise that the control gain does not depend on global information. Finally, two simulation examples are used to verify the feasibility and effectiveness of the proposed algorithms.