Concrete gravity dams are prone to structural safety issues under strong earthquakes, therefore reasonable and economic engineering measures are needed for the protection. A sand retaining gravity dam in northwest China was taken as the research object, and its natural vibration characteristics and dynamic response were analyzed based on the mode-superposition response spectrum method. Taking the tensile stress on the dam heel and the variable cost of seismic measures into account, this study employed the improved multi-objective cuckoo search (MOCS) algorithm for the optimization of anti-seismic design. In order to improve the performance of the algorithm, improvements were made by adjusting the crowding distance calculation formula, introducing polynomial mutation operators, dynamic step size and discovery probability. Through comparative experiments with other multi-objective algorithms on different test functions, it has been proven that the improved MOCS algorithm has the advantages of high computational efficiency and uniform set resolution. Finally, based on the Pareto optimal solution set, the rationality of the design scheme for gravity dams was evaluated, and corresponding engineering suggestions were proposed. This study can provide some valuable insights for the structural design of concrete gravity dams in high-intensity earthquake areas.