<p>In steep and complex geological environments, identifying source areas and evaluating the susceptibility of large-scale rockfalls has always been challenging. To improve the accuracy of evaluations from the perspective of unit division, this study takes the rockfalls in Henghe Township, Wanzhou District, Chongqing City, China, as an example. Key factors such as lithology, slope structure, historical hazards, and joint features are comprehensively considered. Based on the Unique-condition Unit method, a Multi-Condition Unit (MCU) method is proposed, adhering to the principle of the geological environment “consistency” for the division of rockfall mapping units. This method employs Unmanned Aerial Vehicle (UAV) oblique photogrammetry technology to obtain structural information on the surface of hazardous rock-belts (stratum attitude and joint characteristics) and rockfall information on the slope. And reconstructs a three-dimensional (3D) model of the cliff using high resolution image data obtained by UAV. The susceptibility of the hazardous rock belt is inferred based on the distribution density of the nearest points of rockfalls on the hazardous rock zone. Through validation of historical rockfall source areas, rockfall nearest point tests, and practicality tests for susceptibility assessment (Receiver Operating Characteristic (ROC) accuracy analysis), the MCU method was compared with the Grid cell (GC) and Unique-condition unit (UCU) division methods, yielding results of 35% (units with historical source area ratio greater than 50%), 68.43%, and 0.92, respectively, all superior to the other two methods. The results indicate that compared to UCU and GC, MCU not only accurately considers the special geological environmental factors of rockfall development in steep multi-level cliff environments but also helps better identify rockfall source areas.</p>