This paper considers the stabilisation control problem of upper-triangular nonlinear time-delay systems. One distinct characteristic of this work is that the systems contain unmodelled dynamics, unknown control coefficients and time-varying delay, which make the control design much more difficult. By modifying the homogeneous domination approach and introducing a gain-scaling method, a new controller is constructed such that all the states of the closed-loop systems are bounded and convergent to the origin. Simulation examples are given to illustrate the validity of the theory.