Japan, like many countries, is facing problems with an aging society, and lifelong learning is becoming more and more important. To provide older adults with the opportunity to enroll in lifelong learning programs, it is essential to offer suitable programs. However, designing learning programs for older adults is not easy because they may have various physical and mental conditions. To approach the problem, the authors have proposed a mechanism for personalizing learning programs for older adults. The basic ideas of the mechanism are: (1) omitting unessential steps and (2) suggesting collaboration with others. In this paper, the algorithm and the data structures of the mechanism are explained, and results of experiments are shown.