Current neural network-based approaches for redundant manipulator control often suffer from the high computational burden and insufficient consideration of orientation tracking. For instance, methods discussed in earlier chapters do not incorporate orientation tracking in the motion control of redundant manipulators. To address these limitations, this chapter introduces a novel multi-criteria control framework supported by a training-free dynamic neural network (DNN), which concurrently accounts for orientation-tracking constraints and physical limitations. In addition, compared to existing approaches tackling the same task, the presented DNN model exhibits the lower computational complexity. Theoretical analysis verifies that the presented scheme utilizing the DNN model achieves the global and exponential convergence to the theoretical solution for robotic motion control. Moreover, illustrative simulations and real-world experiments conducted on the Franka manipulator further validate the effectiveness and practicality of the presented DNN model.