The random walk model (RWM) can accurately predict the jamming probability of frictionless granular system in a silo. However, the applicability of this theoretical model for systems with friction remains unclear. In this study, a measuring device based on array CCD cameras is built. The characteristic parameters of jamming arch are measured by image method. It is found that the random walk model can accurately predict the silo jamming probability when the mean angle between two adjacent granules is negatively linear with the granular position on the jamming arch and the mass discharge rate is greater than 8.3 g/s. The experimental results extend the application scope of the RWM and provide a reference for the control process of silo devices in industrial production.