This chapter focuses on the key technology of autonomous robot grasping. It begins by analyzing traditional grasp detection methods based on analytical and empirical approaches, highlighting their limitations in real-world applications. With advancements in deep learning and computer vision, deep learning-based grasp pose detection algorithms have become widely adopted. These algorithms leverage RGB-D sensor data, enhancing generalization for grasping unknown objects. While these methods perform well in terms of accuracy and adaptability in practical scenarios, they still face challenges related to real-time performance and the need for large-scale annotated datasets. The chapter concludes by presenting technical case studies that demonstrate the comparative performance of grasping algorithms in real-world applications.