In order to weaken the influences of system uncertainties and coupling items on the coordinated tracking control performance of the speed and tension system of the reversible cold strip rolling mill, an adaptive nonsingular terminal sliding mode (ANTSM) backstepping control strategy using disturbance observers is proposed. First, time-varying gain extended state observers (TVGESOs) are constructed to dynamically observe the system's mismatched uncertainties, which weaken the initial peak phenomenon of the traditional extended state observer (ESO), and effectively improve the system's tracking control precision. Next, based on the backstepping control and the second order sliding mode integral filters, adaptive nonsingular terminal sliding mode controllers (ANTSMCs) for the speed and tension system of the reversible cold strip rolling mill are designed, which solve the “differential explosion” problem of conventional backstepping control, and make sliding mode variables converge in finite time. Again, neural network adaptive method is used to approximate the system's matched uncertainties, and the approximation values are introduced into the designed controllers for compensations. Finally, simulation research is carried out on the speed and tension system of a 1422 mm reversible cold strip rolling mill by using actual data, and the results show the validity of the proposed control strategy.