Drum is one of the important load-carrying components of cranes. The drum’s design,reasonable or not,is of great importance to the security of cranes. Multilayer drum is the current mainstream trend of the development of the drum. It’s significantly to explore the lightweight design. This paper takes a model of multilayer drum of tower crane for example. It sets up the APDL parameterized model according to the Ansys,a large universal software,and analyzes the strength and stiffness of the model. It takes the strength and stiffness as constraints based on the analysis and sets up the optimization design model that aims at the lightest weight. And the lightweight design of the drum comes true under the Design Opt module. In order to reduce the execution cycle of the ANSYS optimization method and improve the efficiency,it uses LM-BP neural network which has great nonlinear problem fitting ability and GA which has excellent global search ability in Matlab to realize this rapid optimization design of multilayer drum. Two discussion,which aim at lightweight design of the drum involved in this paper,have some reference value for the structural design and optimization of multilayer drum of cranes in the future.