When a manipulator executes a task involving object interactions, like polishing, relying solely on the precise motion control is insufficient to guarantee task success, as discussed in Chapters 2 through 5. Given the problem, this chapter presents an acceleration-level motion/force control (AMFC) scheme that incorporates a motion/force-based cyclic motion planning (MFCMP) strategy and an improved fuzzy dynamic neural network (FDNN) model. Specifically, the MFCMP strategy incorporates four essential metrics: minimum joint error, orientation maintenance, motion/force control, and joint physical constraints. In addition, the improved FDNN model is constructed with automatic adjustment on the convergent parameter, which is noise tolerant with the aid of dynamic variables. Furthermore, theoretical analyses are conducted to validate the convergence and robustness of the presented scheme. Comparative results under both noise-free and noisy conditions are then provided to demonstrate the effectiveness, robustness, and superiority of the AMFC scheme.