Conventional research thinks that when vegetation cover is less than 40%, sandland would be in a state of semi-mobile or semi-fixed. However, in practice it has been found that at low vegetation coverage, the horizontal distribution pattern of shrubs can have marked effects on fixation of shifting sands. In this study we investigated the windbreak effects of a Caragana korshinskili kom plantation with 20%25% cover established for the purpose of sand-fixation. Wind velocity was measured at different locations and heights on sites of the intact or gapped belt scheme plantation and randomly distributed plantation by using multiple-sensor automatic anemoscope (GB-228) during windy season (February to April). Statistical analysis indicated that when cover was at 20%25%, the belt scheme C. korshinskili plantation was 48.4%, 30.7%, and 27.4% more effective for wind breaking than the randomly distributed plantation at 20 cm, 50 cm, and 200 cm above the ground surface. The effectiveness of wind breaking increased with wind velocity in the belt scheme. When the wind velocity at 200 cm reached 67 m/s, the mean windbreak effect of the belt scheme was on average 48.2% more than the randomly distributed plantation scheme at the same vegetation coverage. This shows that when vegetation cover is sparse, the horizontal distribution pattern of shrubs becomes an important factor for restricting windbreak effect, and that the belt scheme is the most effective plantation in the windbreak effect. The “narrow pipe” flow from gaps between shrubs can have an effect on increasing wind velocity, which results in 41.3% of observations displaying wind velocity greater than that measured in the open. Within the belt scheme C. korshinskili plantation, none of the observations was greater than observations made in the open ground. The flow field structure of the randomly distributed C. korshinskili plantation was complex and changeable, and was an important factor for the occurrence of wind erosion and sand accumulation in the thin stands and increased wind velocity in the gaps of stands low in coverage in the semi-mobile or semi-fixed sand-dunes. Soil surface roughness was found to be about 5.4114.4 times greater in the belt scheme than in the randomly distributed plantation, indicating that the belt scheme had marked effect on preventing wind erosion and fixing drift sand. When there were gaps in the belt scheme plantation, the reduced wind velocity apparently displayed accumulation phenomena and increased wind velocity was limited to a certain extent in gaps. This study provided scientific basis for developing low cover and belt scheme shrub plantations for sand-fixation.