There are three-stage degradation processes for aircraft structures which are normal stage, crack stage, and after-repaired-stage respectively. However, there has been limited maintenance model for three-stage degradation data of aircraft structures. Therefore, this paper proposes a new approach for three-stage maintenance optimization management. Based on non-linear Wiener process and three-stage degradation analysis, a novel maintenance cost method is presented for aircraft structures and a novel four-level planning algorithm is developed to solve the maintenance cost problem. The first level planning aims at acquiring maximum usage time and the other three-level planning minimizes the maintenance cost of three stages respectively. The experimental data from China Eastern Airlines is adopted to validate and the experimental results show that the proposed method effectively expresses three-stage degradation characteristics to decrease the total maintenance cost; the proposed algorithm achieves a better optimal maintenance cost rate and saves runtime by 10%-20% than the state-of-the-art methods.