An adaptive robust control combined with a multi-objective parameter optimization method for the parallel robot with unknown uncertainty is proposed. In the active joint space, the accurate dynamic model of the parallel robot can be obtained by combining the closed-chain constraint force and the open-chain system's dynamic equation. According to the Udwadia-Kalaba theory, the closed-chain constraint force imposed by the end effector can be calculated in a simple way. The proposed adaptive robust control could guarantee deterministic robust performances of the system (the uniform boundedness and uniform ultimate boundedness). To seek suitable weighting factors for the proposed control, a system performance function, which includes the transient performance portion, the steady state performance portion, and the control cost, is introduced. By applying $D$-operation, the performance function is transformed into a multi-objective function with weighting factors. Meanwhile, the problem of choosing the optimal gain is equivalent to the problem of finding the minimum value of the system performance index function. An illustrative example illustrates the superiority of the proposed modeling method and the proposed adaptive robust control.