Soil microbial biomass and enzyme activity are sensitive to environmental changes and disturbances; consequently, both parameters have been widely used as indicators of ecosystem responses to vegetation succession. However, there remains no consensus about how these variables change during vegetation succession. This knowledge gap may exist because previous studies usually adopt classical statistical analyses, which are not able to reflect the direct influence of one variable on another. However, with path analysis, direct path coefficients and indirect coefficients may be realized, which allows direct and indirect relationship between multiple independent variables and dependent variable to also be obtained. In the present study, the dynamics and influencing factors of soil microbial biomass and enzyme activity during vegetation restoration were investigated. Soil samples were collected from the upper soil layer (0-15 cm) at four vegetation succession stages (i.e., herb, shrub, secondary forest, and primary forest) in a typical karst peak-cluster depression area in northwest Guangxi, China. The characteristics of soil microbial biomass and soil enzyme activity were analyzed with classical statistics, while the one-to-one relationships of soil microbial biomass, soil enzyme activity, soil physical properties, and soil nutrients were obtained by path analysis. Overall, soil microbial biomass carbon (SMBC), soil microbial biomass nitrogen (SMBN), and alkaline phosphatase (ALP) activity increased with vegetation succession. Sucrase (SAC) activity was ordered: secondary forest > herb ≈ shrub > primary forest. Urease (URE) activity was ordered: herb ≈ secondary forest ≈ shrub > primary forest. Path analysis revealed that the direct and main influencing factor for soil microbial biomass was soil organic carbon (SOC). The main influencing factors for SAC were SOC and SMBC, while the total path coefficients of all the factors were small. The main influencing factor for both URE and ALP activity was total nitrogen (TN). TN had a major negative effect on URE activity, whereas it had a major positive effect on ALP activity. Furthermore, the residual path coefficients of SMBC and SMBN, in addition to SAC, URE, and ALP activity, were large, implying that other factors had been neglected in the analysis that have a strong influence on the dynamics of soil microbial biomass and enzyme activity.