A finite-time tracking control scheme is proposed in this paper based on the terminal sliding mode principle for motor servo systems with unknown nonlinear dead-zone inputs. By using the differential mean value theorem, the dead-zone is represented as a time-varying system and thus the inverse compensation approach is avoided. Then, an indirect terminal sliding mode control (ITSMC) is developed to guarantee the finite-time convergence of the tracking error and to overcome the singularity problem in the traditional terminal sliding mode control. In the proposed controller design, the unknown nonlinearity of the system is approximated by a simple sigmoid neural network, and the approximation error is diminished by employing a robust term. Comparative experiments on a turntable servo system are conducted to show the superior performance of the proposed method.