Abstract: Soil enzyme activities are sensitive indicators of soil quality. However, their spatial hetero-geneity and the influencing factors are not well known. In this study, geostatistics, Kriging method, one-way ANOVA and correlation analyses were used to examine the spatial variability and influencing factors of activities of six soil hydrolases: β-1,4-glucosidase (βG), β-1,4-xylosidase (βX), cellobiohydrolase (CBH), β-1,4-N-acetylglucosaminidase (NAG), leucine amino peptidase (LAP), and acid phosphatase (AP) in 0-10 cm soil layers in a karst catchment in Guizhou Pro-vince, China. The results showed that the activities of those soil hydrolase had different spatial hete-rogeneity. The optimal models were the spherical models for βX, CBH and AP, the Gaussian model for βG and NAG, and the exponential model for LAP, respectively. The spatial structure ratios C/(C₀ +C) of βG, βX, CBH, NAG and LAP activities were 99.9%, 99.9%, 99.9%, 76.3% and 96.6%, respectively, implying strong spatial autocorrelation and weak influence from topographic factors. The spatial structure ratio of AP activity was 50.0%, suggesting moderate spatial autocorrelation and vulnerable to the influences of random factors. The variation ranges of the hydrolase acti-vities were greater than the sampling distance, indicating that the sampling methods could mirror spatial variability of the soil enzyme activities at a small watershed scale. Land use type significantly affected the activities of NAG and AP. Slope position significantly affected AP activity. The activity of AP was negatively correlated with soil pH, while the activities of other enzymes except NAG were positively correlated with soil pH. Our results provided useful information on the spatial distribution mechanism of soil hydrolase.