Due to the explosive growth of the amount of Web information, the effectiveness of keyword-based searching methods appears to reach a limit. One major reason is that the mixture of content and presentation information hinders machines in understanding the context of Web information and as a result, the performance of the existing search approaches degenerates. To address this challenge, Tim Berners-Lee of W3C envisioned the Semantic Web. In the Semantic Web, the meaning (semantics) of each term of Web information is defined based on ontologies; thus machines are able to retrieve information that is semantically associated with resources containing input keywords. In this paper, we propose the semantic association search system (SASS), which takes into account the associations between resources (web pages) as well as keywords. To achieve this goal, we first created metrics to evaluate the relative importance of each association between resources, and then developed an exploration mechanism based on the spreading activation paradigm to follow the relevant paths of such associations. Demonstrative cases were tested to validate our approach, and the results showed the effectiveness and great potential of our approach.