The present work deals with fractal-based texture measurement of synthetic aperture radar (SAR) images. Various land classes have been identified by classifying SAR images, and the fractal dimension is estimated for differentiating the classes. Three types of SAR images, viz., synthetic images, simulated SAR images, and satellite SAR images, have been used in the study. An issue with fractal dimension-based measurement is how many digits after decimal point should be considered for estimating fractal dimension? Since fractal dimension is calculated as a ratio, the major challenge is the decidability of significant digits for land classes. The issue has been overcome by discussing the feasibility of fractal dimension for the land classes defined by USGS and the calculated ratio for 200×200 values. As an achievement of the work, it is concluded that two digits of the decimal are sufficient to represent the fractal dimension of land classes.