Large deep-seated landslides within active fault zones, characterized by complex structural deformations and unique geotechnical structures, often lead to severe disasters. This study examines the Jinsha River Fault Zone in the eastern Tibetan Plateau, specifically within the Renniang River watershed. Using InSAR, UAV LiDAR, optical remote sensing, geophysical exploration, and drilling investigations, landslide characteristics and deformation mechanisms were identified. The study result revealed 27 large landslides, further subdivided into 50 secondary landslides, with 92% distributed between two branches of the Jinsha River Fault Zone. Landslide density peaked at 7.3 landslides/km
2, over three times higher than non-fault zones. A method of the surface displacement index (Δ
LSDI) was proposed as a discriminative indicator for deformation zones, suggesting significant deformation when Δ
LSDI exceeds 50ppm. Based on InSAR surface deformation velocity, the landslide deformation intensity was categorized into four levels, of which the strong deformation zones (40≤V
LOS≤70mm/y) and extremely strong deformation zones (V
LOS>70mm/y). The Gonghuo deep-seated landslide, developed two sliding zones at depths of 45 m and 48.85 m, primarily composed of low-resistance black gravelly clay. The deformation exhibits a periodic pattern of creep-accelerated creep-stabilization-re-creep, corresponding to rainfall patterns. The rainfall influence lag-effect for landslide deformation lasts 5-7 days under heavy rainfall, and 10-15 days under moderate rainfall, with acceleration periods lasting 1-2 months to reach the stable stage. The primary deformation pattern of large deep-seated landslides in the active fault zone involves river erosion - traction at the front edge - tensile deformation in the middle resulting in slow-moving. Understanding the deformation patterns and rainfall influence lag-effect in large deep-seated landslides is crucial for disaster mechanism studies and risk mitigation of slow-moving landslides.