A new method of three dimensional (3-D) analysis by Auger electron spectroscopy (AES) is proposed. The 3-D data for a specimen surface layer are compiled by repeating Auger image acquisition and ion sputtering alternatively. Three different techniques for extracting revealing hidden information from the compiled data are shown. 1) Cross-sectional images were constructed by looking at the intensity distribution along a line for each of the stored images. 2) The selected spaces (e. g. grain boundaries) of the 3-D data were quantified. An binary image generated by processing a secondary electron image simultaneously acquired were used to cut out the desired data space as a mask. 3) Depth profiles for the specified area were also constructed. Although 3-D data acquisition takes a long time, the data processed over a wide space shows high reliability, because it is substantially accumulated a large number of data.