ROCCO ring-net flexible barriers play a crucial role in mitigating granular flows induced by landslides in steep mountainous regions. In geotechnical engineering practice, the design of these barriers critically depends on two key factors: the maximum jump height of the granular flow and its peak impact force. While ring-net flexible barriers are known for their deformability and permeability, these characteristics remain poorly understood from a quantitative perspective. To further reveal the impact-jump mechanisms of granular flows against ROCCO ring-net flexible barriers, an array of small-scale laboratory flume experiments were conducted. To modify the permeability of the barrier, three groups of particles with different median diameters were configured to control the relative diameter ratios between the ring-net mesh size and the grains from 2.0 to 3.6. The flow depth and velocity of the incoming granular flow were adjusted by altering the channel inclination to ensure the Froude number between 3 and 10 for dynamic similarity. Specifically accounting for barrier deformation and material outflow, the semi-empirical analytical models, grounded in the principles of momentum and mass conservation, were established. Futhermore, the proposed models were validated by comparing the normalized jump height, and the impact force coefficient at the moment of peak impact force between the prediction value and the experiment data. The experimental results show that both the incoming flow characteristics and the relative diameter ratio lambda jointly determine the impact-jump mechanisms: pile-up or run-up. A larger lambda tends to transition the impact-jump mechanism from pile-up to run-up under the flow conditions with a high Froude number Fr, while the corresponding maximum granular jump height and peak impact force decrease as expected. Comparison between the proposed models and experimental results indicates that barrier deflection determines the upper limit of the jump height, while the lower limit is further controlled by the outflow mass flux. The improved hydro-dynamic impact force model can adequately address most run-up scenarios, whereas, for pile-up cases, the contribution of the hydro-static force should also be considered.