Abstract: The data-driven optimal tracking controls (OTC) for the unknown multi-input system are proposed in this paper, and a novel tuning law is used to update NN weights in the learning scheme.