[Objective] The morphological spatial distribution characteristics of landscape patterns at peak-cluster depressions in rocky desertification areas were determined in order to provide theoretical reference and guidance for revealing the dynamic changes in landscape pattern and ecological protection of peak-cluster depression rocky desertification areas. [Methods] Morphological spatial pattern analysis (MSPA) was used to identify, process, and classify rocky desertification landscapes in order to obtain seven non-overlapping landscape types such as core and loop lines. Landscape dynamics, landscape pattern index, landscape transfer matrix, and hotspot analysis were applied to determine the morphological spatiotemporal evolution characteristics of landscape patterns in peak-cluster depression rocky desertification areas. [Results] ① In 2000, the morphological landscape types were the most widely distributed (190.60 km²), and in 2022, they were the least distributed (147.32 km²). Among the morphological landscape types, the core was the main landscape type in the study area, with a maximum area of 121.62 km² in 2000 and a minimum area of 76.05 km² in 2022, mainly located in the northwest and southern regions of the study area. The area of isolated islands was the smallest among the morphological landscape types, with a minimum of 1.12 km² in 1990 and a maximum of 3.07 km² in 2022. Landscape types such as isolated islands and branches were scattered among various cores. ② The spatial distribution of morphological landscapes in the research area tended to be dispersed, and the diversity, complexity, and fragmentation of morphological landscapes increased. During the research period, the single dynamic degrees of morphological landscape types were 0.10, -0.18, and -2.13, respectively, indicating a shrinking trend in morphological landscape area. The development of morphological landscape types was most intense from 1990 to 2000, and the development was most gentle from 2000 to 2010. ③ The main direction of morphological landscape type transfer was the transfer of the core landscape type to the edge and pore landscape types. Overall, the main direction of morphological landscape transfer was the transfer of the morphological landscape to the background, and the amount of background transfer in was greater than the amount of transfer out. High-high agglomeration areas showed a trend of expansion towards Runhe, Laojie, and Anle Village, while low-low agglomeration areas showed a trend of expansion towards Sanguang and Laojie Village. [Conclusion] The main monitoring indicators for disaster chain early warning include cumulative rainfall, duration of rainfall, changes in moisture content, and on-site monitoring of slope deformation development. These monitoring parameters are influenced by various factors such as rainfall conditions, soil properties, slope angle, and external dynamic conditions. Therefore, in practical applications, it is necessary to comprehensively consider these factors.