<p>Current improvement of negative pressure adsorption system for climbing robots mainly focuses on optimizing the centrifugal impeller, but less attention is paid to the negative pressure cavity and sealing mechanism. In addition, the traditional single-discipline optimization method disregards the interaction between fluid and structure. This research proposes an optimization method using fluid-structure interaction (FSI) and non-dominated sorting genetic algorithm III (NSGA-III) to improve aerodynamic and structural performance of the system, which fully considers the matching relationship among three components and the interaction between fluid and structure. This method was implemented in the optimization of a negative pressure adsorption system in a case study. The gap height, cavity radius, blade number and blade outlet angle were identified as optimization variables with aim to optimize the adsorption force, impeller efficiency and maximum deformation. Results show that the adsorption force increased by 65.8 %, impeller efficiency increased by 4.7 %, and maximum impeller deformation decreased by 38.2 %.</p>