Abstract:Machine learning is the core of artificial intelligence (AI) and one of the fundamental ways to make computers intelligent. With breakthroughs in machine learning algorithms represented by deep learning, artificial intelligence has shown a trend of accelerated development and has been widely used in various industries. Machine learning has significant advantages in computational efficiency, accuracy, portability, coordination, flexibility, and ease of use. The next step will be to change the traditional weather observation mode and accelerate, and improve the processing of weather observation data and the quality of numerical weather prediction and promote the cross integration of earth sciences. In order to better promote the application of AIrelated technologies in meteorology, this paper summarizes the application status of machine learning algorithms from the aspects of meteorological observation, numerical forecast, dangerous weather identification and early warning and satellite data processing.