Abstract: Some disadvantages in NSGA-algorithm Ⅱ are found throngh multi-objective optimization for Pareto-optimal solutions. Three improved techniques are proposed in this paper for accelerating the convergence speed, enlarging the population diversity, and enhancing the uniformity of spread of solutions of the NSGA-Ⅱ algorithm. The three techniques are (1) sorting strategy with the accumulated fitness, (2) disconnected filling algorithm within the established threshold based on the elitism strategy, and (3) dropping strategy with the given threshold. The simulations prove that the improved algorithm has much better convergence than the traditional NSGA-algorithm. Simultaneously, a better Ⅱ optimization result of the linear antenna array patterns can be obtained as well by using this algorithm.