Abstract: With the rapid development of science and technology, the coal mine boom-type roadheader is developing towards the direction of automation, fewer people, and more intelligence. The formation of roadway sections is the main task of the boom-type roadheader. A non-singular terminal sliding mode control method based on an unknown dynamic estimator (UDE) is proposed aiming at the trajectory control of the cutting head for the coal mine tunneling robot, which is affected by unknown external disturbance, system nonlinearity, and parametric uncertainty. Firstly, considering the hydraulic driving system, a 3-order nonlinear strict feedback state-space model of the roadheader cutting electro-hydraulic driving system is established. Secondly, a super spiral sliding mode differentiator with finite time state convergence is used to estimate the unknown speed state of the system. The UDE is designed based on a low-pass filter to estimate the lumped uncertainty of the unmatched items of the system, and a fast terminal sliding mode control law is designed using a nonsingular terminal sliding mode surface and a power exponential reaching law. Real-time feedforward compensation for the system disturbances observed by the UDE can effectively alleviate the controller chattering problem existing in traditional sliding mode control. At the same time, in order to reduce the computational complexity of derivatives of high-order virtual control variables, the command filtering technology is introduced, and high gain robust feedback is used to suppress the disturbance of system matching terms. The stability of the closed-loop system is proved by using the Lyapunov synthesis method. Finally, the prototype of the tunneling robot is utilized to carry out experimental verification. The experimental results show that when tracking step signals, 0.1 Hz sinusoidal signals, and mixed frequency signals, the average value of the tracking error of the proposed control method is reduced by 61.60%, 66.51%, and 66.27% respectively compared with the traditional PI control. And the root mean square value of the tracking error is reduced by 54.00%, 64.53%, and 61.90% respectively. The algorithm can effectively suppress the multi-source uncertainty disturbance of the system, and realize the high-precision position control and dynamic fast response of the cutting system of the roadheader robot.