[Objective] The difference of soil infiltration performance at different altitudes and its influencing factors in Hengshan Mountains were studied to provide ideas and methods for the reseraches of large-scale soil infiltration.[Methods] Using point-source infiltration method test at 9 points in different altitude areas of Hengshan Mountains.[Results] From the point of view of infiltration rate, the infiltration rate changed significantly and the stable infiltration rate increased gradually with the rising of elevation (top > hillside > foothill). From the beginning of infiltration to stable infiltration, the infiltration rate decreased by 50%~86%, and the infiltration process showed a certain regularity. The initial infiltration rate was very large. With the process of infiltration, the infiltration rate decreased rapidly in the first 10 minutes and continued to decrease slowly in the succedent 10~30 minutes. Finally, it got stable after 30 minutes. There are many factors affecting soil infiltration performance. In this experiment, the stable infiltration rate had a significant positive correlation with the content of>0.25 mm water-stable aggregates and with the altitude, but had a low correlation with organic matter, soil bulk density and texture. Kostiakov model, Philip model and Horton model all had more than 0.95 determinant coefficients R² in each test site, which can fit the infiltration rate of Hengshan Mountains test site well. The parameters A of Philip model were negative, meaning that their physical meaning is not clear. Therefore, that model has some limitations in fitting the infiltration rate of Hengshan Mountains.[Conclusion] Horton model can reflect the characteristics of stable infiltration rate at different altitudes (top > mountainside > foothill). Kostiakov model can better reflect the variation characteristics of infiltration rate (foothill > top > mountainside). It also shows that point-source infiltration method is feasible to measure mountainous infiltration.