Abstract: Damage detection is the most difficult part of Structural health monitoring systems. In order to acquire response information as much as possible, sensors should be carefully placed. Taking Tsing Ma bridge as an example and using the genetic algorithms (GAs), this paper discussed the optimal placement of sensors on suspension bridges for detection of accumulated damage for structural health monitoring systems. After a brief discussion on the classical GAs and background of the monitoring systems, the paper described the generalized GAs, and used an example to compare them with the clasaical GAs. The results show that the generallied GAs are clearly more effective than the classical GAs. On the baals of displacement and curvature modes, three expressions of the fitness were defined for different sensors and tasks. The generalized GAs were used to search the optimal placement of accelerometers and strain gages on the stiffening decks and the towers of Tsing Ma bridge. The results show that the generalized GAs give reliable and stable results with rapid convergence.