In order to comprehensively evaluate the effects of revetment vegetation on soil anti-erodibility, the soils under grasslands, shrubs, arbor-grasslands and arbor-shrub-grasslands were selected, the effects of different physiochemical indexes of soil anti-erodibility were analyzed by principal component method, and gray relevant analysis was used to study soil physiochemical indexes and their relation with soil anti-erodibility indexes. Furthermore, a coupling model was constructed to comprehensively evaluate the influence of different revetment vegetation patterns on soil anti-erodibility. The results showed that the main factors of soil anti-erodibility were >0.25 mm water stable aggregate, bulk density, organic matter, alkaline nitrogen, available phosphorus, available potassium, mean weight diameter and water stable index, and >0.25 mm water stable aggregate had the greatest impact on soil anti-erodibility. There was a moderate correlation between mean weight diameter, water stability index and soil physiochemical properties. The correlation from high to low was 0.25mm aggregates > organic matter > alkaline nitrogen > available phosphorus > available potassium > bulk density. The system coupling degree between soil anti-erodibility and physiochemical indexes in different revetment vegetation patterns was weakly coordinated, and the coordination from high to low were arbors-grasslands > arbor-shrub-grasslands > grasslands > shrubs > natural slope. The coordination of arbor-grasslands was moderate while and the others were weak. Furthermore, the result showed that arbor-grasslands were the best combination model of revetment vegetation. The coordination degree between soil anti-erodibility and soil physiochemical indexes could provide scientific references for the improvement of soil structure and anti-erodibility as well as for the preferential selection of revetment vegetation.