Agent system that supports needs of travelers is a classic example of an application of personal agents. This paper re-evaluates assumptions behind the design and implementation of an agent-semantic travel support system (TSS) that was proposed in 2006. The goal is to modify these assumptions to match them with the current user-demands and adjust to the technological changes. On the basis of this analysis, the design of a new TSS is introduced. The key element of the proposed system is semantic matchmaking, which is used to establish, which travel services are to be suggested to the user (an in which order). Therefore, we describe in detail the proposed matchmaking algorithm and its implementation, and illustrate its work through selected examples.