Abstarct: In this paper, a new gradient-baxsed neural network approach is proposed for solving nonlinear programming problems (NLPPs), bi-objective optimization problems (BOOPs), multi-objective programming problem (MOPP) and multi-level programming problem. The main idea is to convert the NLPPs and the BOOPs into an equivalent optimization problem by the mean of the weighted sum method, the MOPP reference point, the MLPP into a to single level problem by the Kuhn–Tucker conditions. where the Pareto optimal solutions are obtained by using different weights. Also the decomposition of parametric space for BOOP is analyzed in details baxsed on the stability set of the first kind. Finally A neural network approach is then constructed for solving the obtained convex programming problem. baxsed on employing Lyapunov theory, the proposed neural net- work approach is stable in the sense of Lyapunov and it is globally convergent to an exact optimal solution of the MOPP and the MLPP. the most prominent feature of the proposed approach is that it can converge rapidly to the equilibrium point (optimal solution), for an arbitrary initial point.The simulation results also demonstrate that the proposed neural network is feasible and efficient.