Cloud Computing paradigm continues growing very quickly. The underlying computational infrastructure has to cope with this increase on the demand and the high number of end-users. To do so, platforms usually use mathematical models to allocate the computational resource among the offered services to the end-user. Although these mathematical models are valid and they are widely extended, they can be improved by means of use intelligent techniques. Thus, this study proposes an innovative approach based on an agent-based system that integrated a case-based reasoning system. This system is able to dynamically allocate resources over a Cloud Computing platform.