Abstract: In the early identification of geological disasters such as landslides and collapses, multi-source remote sensing technologies based on InSAR, optical remote sensing, and airborne LiDAR have played an important role. In view of some problems in the application, based on the practice of many projects, the authors in this paper summarized the applicable conditions of multi-source remote sensing technology in the field of early identification of geohazards in mountainous areas of southwestern China, and put forward corresponding application suggestions. The main understandings are as follows: ① The InSAR identification effect of potential geohazards is restricted by vegetation coverage, terrain conditions and data source types. And it should be paid attention to the chosen of key data processing parameters in the application of wide-area InSAR. Besides, it is necessary to establish the field judgment methods for the InSAR identification as soon as possible, and unify the evaluation criteria of the “deformation degree” of the surface deformation area and the InSAR identification. ② The early identification of geohazards by optical remote sensing is suitable for the three types, including slope area with obvious deformation characteristics, historical slope geohazard, debris flow and potential debris flow ditch. Appropriate data sources should be chosen according to the requirements of different work tasks. And the analysis of the background conditions of the geological environment should be paid attention, while the shortcoming of “micro-deformation” detection should not be challenged. ③ Airborne LiDAR has obvious advantages in early identification of geohazards in high-density vegetation coverage areas, comparing with InSAR and optical remote sensing. The laser point cloud density should be no less than 30-50 points/m² in actual applications. ④ Nap-of-the-object photogrammetry technology is suitable for early identification of high-position collapse (dangerous rock mass), which belongs to the category of optical remote sensing, and its identification effect is also affected by the degree of vegetation coverage. The reasonable detection distance should be set in application and specific value of distance can be adjusted according to the regularity of the dangerous rock wall, which is recommended to be no less than 30 meters. ⑤ Suggestions were proposed, including the comprehensive application of multi-source remote sensing technology and methods with complementary advantages, multi-level arrangement of work deployment, focusing on interdisciplinary application and comprehensive training of interpreters and establishment of geohazards information acquisition capacity progress as soon as possible.