Abstract: The spatial distribution and morphological characteristics of landslides can reflect their development degree and provide a basis for landslide prevention in the region. Taking the loess landslide in Zhidan County as the study object, the nearest neighbor index (NNI) and kernel density estimation (KDE) are selected based on GIS to analyze the spatial distribution regularity of landslide. The morphological characteristics of landslide are studied through statistical analysis. The results show that the INNI value of the study area is about 0.177, with aggregated distribution of landslide in space. The maximum KDE value distributes in the slopes on both sides of Zhouhe River in belts with multiple high-density areas. The calculated landslide area density and point density are 0.26% and 0.19 per square kilometer respectively. The slope gradients of landslide are generally above 70°, the aspects mainly concentrated in south and southeast, and the slope heights mostly below 40 m. The landslide in the study area can be classified into three different morphological types by angle, height and geological environment characteristics through cluster analysis.