An adaptive finite-time consensus control method is investigated for nonlinear strict-feedback multi-agent systems that contain unknown parameters and a dynamic leader with both input and output. By using command filters, the “explosion of complexity” phenomenon in traditional backstepping technique can be avoided. The filter errors can be compensated by introducing compensating signals. Adaptive finite-time controllers are constructed by a backstepping technique, command filters, and compensating signals. In addition, the consensus performance and stability of closed-loop systems can be guaranteed in finite time basd on a practical finite-time stability criterion. Finally, a simulation example demonstrates the feasibility and effectiveness of the proposed adaptive finite-time consensus control method.