Abstarct: fragmentation is an important factor for blasting operation successfulness and can reduce costs of a mine. Nowadays for cost reducing and time saving reasons of image processing method, this operation is a routine method at fragmentation operation and optimum pattern recognition of blasting. Several works throughout the world were done to determine the gradation of fragmented rock masses with the use of image processing soft wares. But because of expensive costs of screen analysis and either its non operational conditions, in much cases results were without proper investigations and necessary verifications. In image processing science, some factors are affective to fragmentation investigation. Some of which that can be mention are angle of photography, number of necessary pictures, shape of muck pile, gradation, distribution of fragmented parts, soft ware that used and etc. the goal of this research is to assess the effect of mentioned factors on results of two image processing soft ware, “GOLDSIZE” and “SPLIT”; and also evaluation and comparison of results of both soft ware’s. For this research a pile of fragmented rocks hauled to lab. First the pile was screen analyzed and its curves plotted. In next step, these rocks photographed at three different situations like spread, bench shape and pile. The effects of change in angle at bench situation, on results of image processing, with change of photography angle in different angels were surveyed. Photographed pictures analyzed by “GOLDSIZE” and “SPLIT” soft wares. From experiments were done and analysis of gradation curves it can be seen that spread shape of fragmented rocks have the best gradation curve and also the bench shape of rocks have a better and nearer results to reality in relation to muck shape. The change of photography angle doesn’t have much effect on image processing gradation results. In compare of two mentioned soft wares, it can be seen that the “GOLDSIZE” magnify the results (toward screen analysis) and the “SPLIT” show it finer than screen analysis. In fine particles analysis, the error of “SPLIT” were greater than of other parts of gradation, this is in such manner that results of fine particles are not acceptable. The error of “GOLDSIZE” is less than of “SPLIT” and “GOLDSIZE” is much precise than of “SPLIT”.