Abstract: Particle size and shape significantly influence the crushing behaviors of coarse grained soil. In this study, the particle scanning, particle shape analysis, and single particle compression tests on the gravel soils of high filling project of Baihetan Hydropower Station are carried out. The test results show that the crushing strength of individual particles exists noticeable size effects. The characteristic crushing strength decreases with the particle size. The applicability of the Weibull model in describing the size effects on the crushing strength is then evaluated. The particle shape significantly affects crushing behavior. The particle shape is incorporated into the calculation of the crushing strength, with the shape factor calculated using an artificial neural networks algorithm. A probabilistic model is proposed to predict the crushing strength distribution by incorporating the particle size and shape.