At present, the domestic silage industry has certain limitations in terms of market services. In silage testing, some enterprises or organizations can provide silage testing services, such as testing the nutrient composition and microbial content of silage, etc., but there is still a lack of quality prediction and targeted advice and guidance services. Among the existing testing methods, the traditional wet chemical method for silage fermentation quality testing has been applied earlier and is more classical. The principle is to accurately determine all kinds of components in silage through a series of chemical analysis methods. However, the operation process is cumbersome and requires specialized laboratory equipment and technicians with professional knowledge. It often takes a lot of time, ranging from a few hours to several days, from sample pre-treatment to a complete test report. At the same time, the experimental process requires the use of a variety of chemical reagents, these reagents are not only costly, but also may cause some pollution to the environment. Overall, the wet chemical method is relatively costly for a single test, and for large-scale, high-frequency silage testing needs, it faces challenges in terms of both cost and efficiency. Therefore, we pioneered the combination of machine learning and scientific research big data to create a model SMART-Maize that can accurately predict the quality of silage fermentation and nutrient content at the end of the silage period based on corn silage raw material indicators and fermentation conditions.