This study considers a nonlinear optimization problem used to achieve user equilibrium in the network traffic assignment problem. By providing the Karush Kuhn Tucker conditions of this optimization problem, it is converted into a system of differential equations using the Lagrange function. This system is then redefined as a Lagrange neural network, which is proven to be asymptotically and lyapunov stable. Finally, a numerical method are used to demonstrate that the results obtained from this neural network are a solution to the optimization problem and converge to user equilibrium.