Superfine pulverized coal combustion is a new pulverized coal combustion technology which has better combustion stability, higher combustion efficiency, lower NO x and SO 2 emissions, and higher comprehensive efficiency than when using conventional particle sizes. In this paper, we novelly applied both the classical single power law fractal dimension and a piecewise approach to analyze the particle-size distribution (PSD) of superfine pulverized coal particles. In addition, we introduced the fractal theory into scanning electron microscopy image analysis by adopting the slit island method. The grey relational analysis was used to study the degree of relative importance of the influencing factors about fractal dimensions. Furthermore, the piecewise characteristics of the PSD of coal particles were studied in detail. All curves on the log-log scale can be approximated by two intersecting lines, and each curve can be quantified by four variables. The grey relational analysis was also used for further study on the relationships between the four parameters. Finally, a new method for identifying the economic granule size of pulverized coal particles, that is, economic fineness based on the power consumption of coal mills, E 2 , was proposed by a utilizing neural network method. Final results indicate that the economic fineness of Shenhua pulverized coal particles based on E 2 is about 18.94, while that of Neimenggu pulverized coal particles is about 36.13. This provides some reference for a coal-fired power plant to confirm the economic granule size, which has certain guidance meaning for economical operation and low power consumption.