Jacket-type platforms may be the most common type of offshore structures, and damage localization and severity estimation is important for these structures. This paper employs a multi-hierarchical damage identification method based on BP neural network to detect damages in jacket platforms. Firstly, the damaged storey of the jacket is detected, and the numbers of the elements among the detected storey are then detected. According to this method, the learning samples can be more targeted and the number can be reduced largely. In the end, a jacket model is used to investigate the performance of this method, and the results indicate that this approach is more effective and has higher accuracy than direct diagnosis method.