A semantic caching scheme suitable for wrappers wrapping web sources is presented. Since the web sources have typically weaker querying capabilities than conventional databases, existing semantic caching schemes cannot be applied directly. A seamlessly integrated query translation and capability mapping between the wrappers and web sources in semantic caching is described. In addition, an analysis on the match types between the user's input query and cached queries is presented. Semantic knowledge acquired from the data can be used to avoid unnecessary access to the web sources by transforming the cache miss to the cache hit. A polynomial time algorithm based on the proposed query matching technique is presented to find the best matched query in the cache. Experimental results reveal the effectiveness of the proposed semantic caching scheme.