This paper presents two approaches to implement cost-effective structural health monitoring. The first approach is based on the relation between monitoring cost and time-dependent reliability importance of a structural component. The monitoring cost is allocated to individual components according to their reliability importance factors. This allocated monitoring cost is used to determine the optimum monitoring schedule. The second approach is formulated as the minimization of the expected damage detection delay based on an event tree model. In order to consider the conflicting criteria of minimization of both expected damage detection delay and monitoring cost, a bi-objective optimization problem is formulated and solved.