Soil phosphorus (P) and potassium (K) nutrients are the basis of maintaining the ecosystem function of red beds landform area. Topography is one of the main influencing factors of soil P and K nutrients contents. This study focuses on the soil of typical red beds landform in southern China. The semi variogram model of soil P and K nutrients was established by geostatistics and GIS technology. Moran's I coefficient was used to analyze the spatial autocorrelation of soil P and K nutrients. The spatial distribution characteristics of soil P and K nutrients were studied by the interpolation method, and the correlations between soil P, K nutrients and terrain factors (elevation, slope and aspect) were discussed. The results showed that: (1) the soil total phosphorus (TP) content was relatively high, the soil available potassium (AK) content was at a medium level, the soil TP and soil available phosphorus (AP) content were relatively low, and the variation coefficient in decreasing order was AP (36.39%)>AK (28.43%)>TP (28.23%)>total potassium (TK) (24.33%). (2) The content of soil P and K in the southwest of the study area was higher than other areas, the lowest value in the grassland and the middle value in the red bed bare land. (3) The soil TP, TK, AP, AK showed strong spatial autocorrelation, Moran's I coefficients of soil P and K were all positive, following spatial agglomeration distribution. Moran's I coefficients in decreasing order was AK (0.2174)>TP (0.2000)>AP (0.1897)>TK (0.1665), respectively. (4) The correlation analysis of soil P, K nutrients and terrain factors showed that soil TP, TK, AP, AK nutrients and terrain factors (elevation, aspect) had extremely significant positive correlations (P<0.01), in particular, soil AK has a significant positive correlation with elevation (P<0.01), and the correlation coefficient is 0.901. The results are of great practical significance to soil management and ecological environment management in the red bed area.