The physical and mechanical properties of aggregates are determined by their internal physical matrix-structuring

which cannot be directly demonstrated by traditional gradation curves. Also

there is a lack of optimal target function in designing of graded aggregates used in heavy-haul railway subgrade surface. This study established a grading curve model based on logarithmic probability regression equation and transformation of traditional gradation integral curve

and proposed a coefficient of matrix-structuring ‘R’. The mathematical expression of R was derived and the sensitivity of related parameters was analyzed. Based on the physical meaning of R

a practice engineering approach to optimize graded aggregates used in heavy-haul railway subgrade surface was proposed with the target of optimizing R. The feasibility of the proposed approach was verified by compaction test

large direct shear test and permeability test. The optimal value of R was determined by experimentation and optimization

between 21.17 and 22.08. Findings in this study provide reference for gradation optimization designing of graded aggregates used in heavy-haul railway subgrade surface.