Soil microorganisms have played a vital role in the evolution of organisms, and have an important function in ecological restoration. However, there is much controversy about whether soil microorganisms show spatial distribution patterns. Many forest types have been destroyed in karst regions around the world, and because of the fragile ecological conditions in such regions, it is difficult to restore their vegetation and ecosystems. Here, we conducted a field study to analyze the soil microbial biomass in typical sloping farmland in a karst region on the Yunnan-Guizhou plateau. We selected a typical area of sloping farmland for the field study. The total area was 340 × 320 m, and it was divided into a 20 × 20 m grid using an Electronic Total Station, giving 294 sample points. We examined the spatial patterns of soil microbial biomass using geo-statistical and statistical methods. The soil microbial biomass carbon (C_mic), soil microbial biomass phosphorus (P_mic), and the ratio of soil microbial biomass carbon to nitrogen (C_mic/N_mic) showed moderate values in this karst region on the Yunnan-Guizhou Plateau. The average values for C_mic,P_mic,C_mic/N_mic were 383.80, 15.41, and 5.61 mg/kg, respectively. There was a low soil microbial biomass nitrogen (N_mic) content (71.50 mg/kg). There was a strong correlation between C_mic and N_mic, but not between C_mic and P_mic,or between N_mic and P_mic. All of these factors showed wide variations in magnitude and differences in their spatial distribution patterns. The variation coefficients for these factors ranged from 39.96 to 87.50%, and the smallest Moran's I result for C_mic,N_mic,P_mic, and C_mic/N_mic was -0.088, -0.0539, -0.060, and -0.093, respectively, indicating that the soil microbial biomass showed a high degree of spatial heterogeneity. The best model for all the four variables in the sloping farmland was an exponential model. The RSS values were very small, indicating the good fit of the model. The nugget (C₀) was low (0.0016-0.0087) and all C₀/(C₀+C) values were less than 25%, indicating that the soil microbial biomass showed strong spatial autocorrelation, and that the spatial patterns of microbial biomass were less influenced by random factors. The spatial patterns in microbial biomass showed a small range (22.2-51.0 m). The ranges of the patterns in C_mic and P_mic were similar (51.0 and 50.1 m), and were greater than that of N_mic (22.2 m). This indicated that the four variables showed strong spatial variability, which was caused by structural variability. The Kriging contour maps showed high values for C_mic, N_mic, and C_mic/N_mic on the middle and upper parts of the slope. On these parts of the slope, C_mic showed a clear spatial distribution pattern but N_mic did not. The plaque was thicker and broken and showed higher heterogeneity on the while slope. The spatial distribution pattern of P_mic was more evident on the middle and lower parts of the slope and the slope foot. Therefore, there is not only a small-scale spatial distributional pattern of the microbial biomass on sloping farmland in the karst region on the Yunnan-Guizhou Plateau, but there are also different spatial patterns in soil microbial variables.