Abstract: The geological background of the Huangshui River Basin is complex and diverse, resulting in significant differences in the disaster-causing mechanisms and stability of various landslide types. To reveal the deformation characteristics and stability of loess slopes under different rainfall conditions, this study takes the Lijiamo unstable slope in the Huangshui River Basin as a case study. Using space-air-ground observation technology, field investigations were conducted on the slope. The dual-strength reduction method was employed to simulate the displacement, stress-strain behavior, and safety factors of the slope under natural and heavy rainfall conditions, aiming to analyze and verify its stability under different scenarios. The results show that: ① The Lijiamo unstable slope exhibits a nearly semi-circular shape. Under the combined effects of artificial slope cutting and heavy rainfall, it is prone to transform into a landslide disaster. The failure initiates at the slope foot, which slides first, causing the middle and upper parts to lose support and subsequently slide, characterizing it as a "retrogressive" landslide. ② Currently, the slope is in a stable state. However, due to the development of numerous tensile cracks on the slope surface, local sliding may occur. Nevertheless, heavy rainfall poses a greater risk of causing overall slope instability. ③ From June 2019 to September 2023, the vertical annual average deformation rate of the slope ranged from -4.50 to 1.96 mm/a, and the north-south annual average deformation rate ranged from -0.41 to 0.44 mm/a. From October 2023 to April 2024, the maximum vertical settlement on the slope surface was -9.77 mm, the maximum uplift was 9.77 mm, the maximum north-south settlement was -161.31 mm, and the maximum uplift was 54.94 mm, indicating a stable state during this period. ④ Under natural conditions, the safety factor of the slope is 1.21.Under heavy rainfall conditions, the safety factor decreases from 1.21 to 0.93, indicating an unstable state. The horizontal displacement is largest in the middle of the slope, reaching 12.94 cm, and the vertical displacement is largest at the top of the slope, reaching 12.24 cm. These findings demonstrate that combining space-air-ground observation technology with the dual strength reduction method can provide more comprehensive multi-source heterogeneous data support for slope research.