In this paper a real time control technique for a nonlinear discrete time Multi-input Multi-output systems is presented. The proposed technique is based on the combination of PID controller properties and neural networks characteristic of self-learning and flexibility of function presentation. The standard back propagation (BP) algorithm is used to find parameters of the PID neural network controller. The suggested technique modifies the architecture used in SISO to fit with MIMO systems.  The experimental results are presented toward the end of the paper to show the effectiveness of the proposed technique