Abstract: Radio-frequency power amplifiers are key components in communication systems. When operating at high efficiency, their inherent nonlinear characteristics introduce signal distortions that degrade overall communication quality. As a mainstream linearization technique, digital predistortion can effectively compensate for such nonlinear distortions and enhance system performance. However, conventional digital predistortion approaches are primarily designed for static and single operating conditions, making them inadequate for intelligent communication scenarios characterized by diverse operating states, frequent mode switches, and stringent requirements on real-time processing and low complexity. This paper discusses the requirements and challenges of digital predistortion in intelligent communication scenarios, introduces digital predistortion techniques tailored for such scenarios, discusses their advantages and limitations, and outlines potential directions for future development.