© 2015 Lavoisier. This paper discusses the assessment of Web resources trust. This assessment is undermined by the uncertainty that raises due to end-users’ ratings that can be questioned and variations in Web resources performance at run-time. To tackle the first uncertainty a fuzzy-clustering-based credibility model is suggested so that the gap between specific endusers (known as strict) and the current majority is reduced. To deal with the second uncertainty a probabilistic trust approach is proposed so that trust levels for future interactions with Web resources are made available to users. The probabilistic approach is built upon probabilistic databases and a fuzzy-based credibility model. A series of experiments are carried out to vali- date the suggested credibility model and the trust approach.