The plasma-assistance dry reforming of methane products (hydrogen and carbon monoxide) predicted at atmospheric pressure were simulated using artificial neural network. The experimental data required for modeling artificial neural networks was collected in a corona discharge plasma reactor. Effect of process parameters (plasma discharge power and feed flow rate) on the performance of the reaction, for example, methane conversion, and selectivity of the products were studied. A network with feed forward back-propagation algorithm, and Levenberg- Marquardt training function, tangant sigmoid activation function and linear activation function for hidden and output layer, respectivitly were used in this project. For example, artificial neural network model predicted methane conversion 25.12%, selectivity of hydrogen and carbon monoxide 71.15%, 74.85% in 4 w plasma discharge power. The model error values for the conversion of methane, hydrogen and carbon monoxide selectivity were 0.47%, 1.2% and 0.2%, respectively. A combination of genetic algorithms and artificial neural network to achieve optimum operating conditions were used in the reforming process. The results showed that the optimal feed flow rate and plasma discharge power were 175 ml.min-1and 6 watt, respectively. Also, the conversion of methane and hydrogen selectivity were 25.85% and 65.15%, respectively. The very small differences between predicted and experimental values confirm that combined neural network with genetic algorithm model is suitable tools for modeling and optimization of plasma-assistance methane dry reforming process.