The design of nonlinear tracking controller for antagonistic tendon-driven joint has garnered considerable attention, whereas many existing control methodologies are impractical in the real-time applications due to complexity of computations. In this work, a robust adaptive control method for controlling antagonistic tendon-driven joint is mainly studied by combining adaptive control with mapping filtered forwarding technique. To enhance the robustness of the closed-loop systems, the true viscous friction coefficients are not needed to be known in our controller design. Typically, to tackle the problem of “explosion of complexity”, filters are introduced to bridge the virtual controls such that the controller is decomposed into several sub-modules. Mappings and their analytic derivatives are computed by these filters, and the mathematical operations of nonlinearities are greatly simplified. The block diagram of this controller of tendon driven joint is provided, and controller performances are validated through simulations.