Abstract: The deep reduction of iron oxides and the growth of iron particles were conducted during reduction for refractory iron ore. Currently, metallization rate as an evaluation index of reduction can only evaluate reductive degree, but not particle size characteristics of iron particles. For building intact evaluation system of reduction, digital processing technology was used to analyze features of iron particles. Furthermore, formula for calculating the cumulative granularity property curve of iron particles was derived by utilizing the two-dimensional characteristic parameters, which collected and handled from digital images of material after reduction in combination with the spherical feature of iron particles, with which the size characteristics of iron particles under different conditions could be evaluated effectively. The experimental results showed that the separating index of iron powder products follows the changing tendency of iron particle size precisely. Consequently, evaluation method for the deep reduction of iron ore with digital image processing techniques is feasible.