Multi-Agent System (MAS) paradigm has seemed especially appropriate for developing applications in complex domains, as Real-Time environments are. In this situation, an artificial intelligent technique may require tasks which involve searching in large search spaces in a bounded time. However, the main problem of MAS applications over real-time environments is to merge intelligent deliberative techniques with real-time actions in complex and distributed environments. In this settings, agents must make sure that they will be able to perform a service within a specific time before committing themselves to its performance. This paper proposes the use of a Case-Based Reasoning (CBR) technique to cope with this functionality. The technique is tested in an example where the tasks are bounded and need a temporal control.