Weather Elements Nowcasting based on machine learning (WEN) has the characteristics of high release frequency, high time resolution, and complex forecast model based on climate and time. Using Python libraries such as multidimensional tag array, machine learning tools, parallel computing framework aiming at “fast computing”, a testing subsystem is established. It takes the time range covered by the “prediction model” as the statistical test time boundary. It objectively gives the prediction performance, which provides a basis for evaluating model optimization effect and the operation of products.