The application of information technology in the field of biomedicine has become increasingly important over the last several years. This paper presents an intelligent dynamic architecture for knowledge data discovery in biomedical databases. The core of the system is a type of agent that integrates a novel strategy based on a case-based planning mechanism for automatic reorganization. This agent proposes a new reasoning agent model, where the complex processes are modeled as external services. The agents act as coordinators of Web services that implement the four stages of the case-based planning cycle. The multi-agent system has been implemented in a real scenario to classify leukemia patients. The classification strategy includes services to analyze patient’s data, and the results obtained are presented within this paper.