In this paper, a neural network is designed and learned to accommodate the SVPWM switching
method. The selection of SVPWM (digital PWM method) as ANN application was chosen because
it is nonlinear problem and more sophisticated problem multilevel inverters. The ANN will be
given good results in the real time systems. The learning of neural network was backpropagation
method, and the neural network onrput is used to gating the inverter switches, fmally obtained
intelligent inverter that responds to any new parameters another of that used. to learning it: Neural
network design, learning, inverter output, and harmonic analysis are accomplished by using CU
language. The obtained results are verified by comparison with SVPWM inverter. The ANN
invmer gives very satisfied results according to the inverter criterions.