Abstract: Based on an artificial neural network inverse dynamic model of robot manipulators,a nonlinear model-based control scheme is presented in this paper.The implementation of the whole scheme assumes no anyknowledge of the system.The simulation tests for the two-link hydraulic robot manipulators indicate that theposition control system under consideration has acceptable good tracking performance,and demostrate thatartificial neural networks have potential of solying identification and control problems of nonlinear systems.