With advances in information and communication technology, interactive multimedia learning systems are widely used to support teaching and learning. However, as human factors vary across users, they may prefer the design of interactive multimedia learning systems differently. To have a deep understanding of the influences of human factors, we apply a data mining approach to the investigation of users' preferences in using interactive multimedia learning systems. More specifically, a clustering technique named K-modes is used to group users' preferences. The results indicate that users' preferences could be divided into four groups where computer experience is a key human factor that influences their preferences. Implications for the development of interactive multimedia learning systems are also discussed. (Contains 4 figures and 3 tables.)