Abstract: To investigate the invulnerability of small-world networks against cascading failure, an improved nonlinear load-capacity model was introduced. Under the rule of local weighted load redistribution, the influence of weight coefficient, capacity parameters and density of the network on invulnerability of small-world networks against cascading failure was studied by numerical simulation. Then, qualitative analysis of invulnerability of small-world networks under the restrictions of cost and performance was presented. The results demonstrate that, invulnerability of small-world networks weakens with the increasing of the weight coefficient. In addition, the spreading of cascading failure of small-world networks with different density slows down when the weight coefficient is smaller, and that of the sensitiveness to network density is different. Moreover, there exists an optimal combination of capacity parameters which ensures that small-world networks reach the strongest invulnerability level. Research findings will provide useful guidance and reference in the optimization design of invulnerability for real complex networks.