This paper presents MKM, a meta-knowledge model to manage knowledge reliability, in order to extend a CBR system so that it can reason on partially reliable, non expert, knowledge from the Web. Knowledge reliability is considered from the point of view of the decision maker using the CBR system. It is captured by the MKM model including notions such as belief, trust, reputation and quality, as well as their relationships and rules to evaluate knowledge reliability. We detail both the model and the associated approach to extend CBR. Given a problem to solve for a specific user, reliability estimation is used to filter knowledge with high reliability as well as to rank the results produced by the CBR system, ensuring the quality of results.