Abstract: Permanent gully is well developed in Jinsha Dry-hot valley region in China. The ravine density of this region can be up to 7.4 km/km2 with the maximum soil erosion modulus 1.64×104 t/(km2·a). Gully erosion seriously threatens the ecological security and social-economic development of the region. Vegetation is one of effective measures to cut down soil erosion. In order to explore the mechanism that vegetation affects overflow movement and promotes gradual stability of active permanent gullies, and find out efficient approaches and methodologies to control the development of permanent gullies in this area, a field scouring experiment was carried out between March and April 2013. The study aimed to investigate the impact of grass belt length on runoff sediment yields and transportation process as well as the spatial and temporal distribution characteristics of sediment content and sediment transport rate. Five experimental plots with grass belt length of 0, 4, 8, 12 and 16 m were constructed on an in-situ gully head in the Gully Erosion and Collapse Experimental Station, Institute of Mountain Hazards and Environment, Chinese Academy of Sciences & the Ministry of Water Resources, Chengdu. Observing cross-sections were set in the gully bed every 4 m as well as an observing cross-section in the drainage area in every plot, and the distance between the gully headwall and the observing cross-section in the drainage area was 1 m. In addition, the width of flow section was measured by steel tap, the depth of runoff was measured by straightedge, and the velocity of overland flow was detected by staining method. Furthermore, sediment samples were collected by particular plastic bottles in every observing cross-section, and oven drying method was utilized to determine the sediment content, and sediment transport rate was calculated by sediment content and flow rate of flow cross-section. The results showed that: 1) Increasing grass belt length could not only effectively reduce the sediment content, but also change the spatio-temporal distribution of sediment content; Runoff sediment content decreased in an exponential function with the flushing time and exhibited an increasing trend along with the experimental plots from the upstream drainage area to downstream gully bed; The fluctuation degree of sediment content over the experimental time went through a decrease process with the increase of grass belt length; As the increase of grass belt length, the fitness of sediment content with experimental time became better, that is to say, the downward trend of sediment content became more and more obvious; On the other hand, although the sediment content underwent an increase process, the increasing trend of sediment content became weaker and weaker along with the experimental plots from the upstream drainage area to downstream gully bed; 2) the grass belt could reduce the sediment transport rate; The sediment transport rate in the experimental plots where the grass belt length was less than 8 m were higher than those in plots where the grass belt length was greater than 8 m; Moreover, when the grass belt length was less than 8 m, sediment transport rate exhibited to increase firstly and then change to sharply decrease and then tended to stable over the whole experimental time; However, runoff sediment transport rate had a weakly decreasing trend over the experimental time when the grass belt length was or greater than 8 m; Furthermore, with the increase of grass belt length, runoff transport rate of the same position observing cross-section in every experimental plots tended to decrease; The mean of runoff sediment transport rate of every observing cross-section showed an increasing trend along with the experimental plots from the upstream drainage area to downstream gully bed, while the specific increasing way of sediment transport rate was influenced by the grass belt length. Comparably, 8 m grass belt may be the optimal length for reducing runoff sediment content.