Type IV creep damage occurs in welded joints of high chromium steel in power plant. The evaluation method of residual life of the damaged component is demanded. One of them is the methods based on the number density of voids that is defined as the number of voids divided by the area of observation. The parameter represents the life fraction of components and is influenced by the stress tri-axiality, but the influence is hidden by the influence of the area and magnification of observed metallographic structure. The error of the parameter is large if the area of observation is not large enough, because the voids do not uniformly distribute. The parameter varies according to the magnification because voids can’t be seen if magnification of observed photograph of metallographic structure is small. The appropriate areas and magnification are required for obtaining the appropriate parameters, for studying the influence of stress tri-axiality and for evaluating the residual life. In this study, the creep tests with the specimens of simulated fine-grain HAZ of Mod.9Cr-1Mo steel were carried out and samples with Type IV creep damage were obtained. The distribution of voids under the uniform stress was measured under the different conditions of magnifications and areas. We obtained the mean and the standard deviation of the distribution and proposed the area of observation appropriate for allowable error of the number density of voids that depends on the magnification.