A hypothesis-based reasoning system handles the knowledge-base including complete (fact) and incomplete (hypothesis) knowledge. The incomplete knowledge plays a crucial role to realize advanced AI functions, such as common-sense, flexible matching, learning, etc. Thus the hypothesis-based reasoning is important as one step toward next generation knowledge-based system. In thls paper, an enhanced knowledge representation is described for the hypothesis-based reasoning system. Then an inductive learning mechanism is presented in the framework of the hypothesis-based reasoning. This learning mechanism enables multiple-concepts formation from given examples. The key technology of the mechanism is a minimum generalization extended to the knowledge including hypotheses. The whole system has been fully implemented as a meta-interpreter on Prolog.