High-rate GNSS has become an important tool in seismological research. Unlike inertial instruments such as seismometers, data processing and noise reduction of high-rate GNSS are the key links in extracting reliable seismic waveforms. This study begins with high-rate GNSS data processing, including the development of a software compatible with various satellite precise products to ensure stable and comparable solutions. Furthermore, by integrating artificial intelligence approaches, we enhance the denoising of high-rate GNSS seismic waveforms, significantly improving their signal-to-noise ratio. These improved waveforms are subsequently applied to earthquake source parameter inversion, demonstrating their utility in seismological analysis.