Service-oriented computing introduces the need for trust management that enhances the security and privacy of component services across organizational boundaries without increasing human workload. In this dissertation, the issues of federating distributed and heterogeneous autonomous trust management systems for service-oriented computing are examined from four different aspects—trust establishment, representation, exchange, and enforcement—and four building blocks are proposed accordingly. Trust establishment is in charge of building and monitoring direct and indirect trust relationships with privacy protection; trust representation includes a mechanism to handle uncertainty in trust information from indirect information and subjective judgment; trust exchange uses a hybrid method to permit interoperation of various types of trust information; trust enforcement provides a translation approach to bridge the gap between high level policies and low level enforcement mechanisms. Then these four building blocks are incorporated into a unified approach for federated trust management for service-oriented architectures. Meanwhile, because the literature provides little guidance regarding evaluation methodology for trust management systems, a new evaluation framework, including a set of requirements and evaluation metrics, is proposed.