In order to solve the problem that the traditional algorithm is difficult to apply directly because the perception system of quadruped robot is vulnerable to motion noise interference, limited perspective and computational resource constraints, the intelligent perception technology system of “goal-ontology-environment” is studied and analyzed. By reviewing the latest research progress of related technologies, this paper analyzes the key technologies and methods of target localization, target tracking, target re-recognition and simultaneous localization and mapping (SLAM). The application characteristics, technical bottlenecks and optimization direction of quadruped robot are discussed. The results show that the analysis injects new vitality into the development of intelligent mobile robot and provides an effective reterence for future progress.