The human-robot interaction torque can affect the control performance of hip exoskeleton robots while it is difficult to install sensors to measure the human-robot interaction torque. In order to address this issue and improve the flexibility, an admittance control method based on an enhanced unknown system dynamics estimator (EUSDE) is proposed for the hip exoskeleton robot in this paper. More specifically, the EUSDE based on two low-pass filters and simple algebra operation is first developed in order to estimate the unknown human-machine interaction torque, where only two filter coefficients need to be adjusted and the robustness against noise is enhanced. Moreover, the proposed EUSDE can be easily embedded into an admittance controller such that the flexibility of the exoskeleton robot is enhanced. Finally, comparative simulation results are conducted to show the efficacy of the admittance controller with the proposed EUSDE for the hip exoskeleton robot.