Abstract: The incomplete data, came into being in the process of data collection, always have some similarities to the complete data in the sample space. This article presents a treatment of the incomplete data based on the decisive tree. On the base of the comparison and analyses of the treatment of the incomplete data, a useful decisive tree is used to deal with such data, and at last, it is testified by an experiment. The experiment shows that the decisive tree can achieve a good result in the treatment of the incomplete data.