Potato starch is a valuable and affordable technical raw material for a number of industries. For selection of plants producing starch with optimal processing properties effective methods for physicochemical parameters evaluation of a large number of starch samples are needed. Thus, variability of phenotypic traits data are important both for fundamental works on identification of genomic loci responsible for a wide range of potato starch characteristics as well as for applied accelerated selection of new varieties for technical use. Estimating the morphology of starch granules by microscopy is one of the most accessible and therefore widespread methods of phenotyping. We developed a four-step approach to the estimation of the geometric parameters of starch granules. It includes an isolation of starch from the tuber (stage1), the preparation of micrographs of starch samples (stage 2), processing and analysis of the images obtained in the freely distributed ImageJ program (stage 3), and the construction of distribution curve for starch granules by geometric parameters (stage 4). It was shown that the starch granules of different varieties and hybrids of potato differ in morphology and can be differentiated by microscopy with obtaining data on the Feret's diameter and the circularity of the particles. Thus, typical values of the Feret's diameter of starch granules of "Alena" and "Nevsky" varieties and 785/8-5 hybrid are 5, 22 and 67 microns, respectively. The distributions on circularity of starch granules of these varieties and the hybrid have only minor differences. Light optical microscopy of starch granules followed by digital image analysis is an affordable, economical, simple and effective approach to phenotyping the varieties and hybrids of potato Solanum tuberosum L. on the physicochemical parameters of starch. The approach may be applied for accelerated analysis of a large number of samples on a limited amount of natural material in the field and countryside economic laboratories.