The process of data preprocessing in data visualization leads to data uncertainty inevitably. Taking voxels in the boundary as our objects, this paper raised the possibility distribution measure based on multi-scale space to calculate the distribution of multiple kinds of data. Meanwhile, information entropy was introduced to quantify their uncertainties, which were then displayed as color gradient clothed on surface of volume data. Experiments have shown that our measures are able to effectively quantify and display the uncertainties of the boundaries.