Abstract: At present, Chinese iron and steel enterprises mainly use the traditional "trial and error method" to develop new steel grades. This traditional "research and imitation" process requires a lot of time and R&D costs. In order to improve R&D efficiency, the use of AI (AI, artificial intelligence) methods to realize new product research and development, optimization of process parameters of existing products, and improvement of product quality and reliability are key technologies that need to be solved urgently by iron and steel enterprises. This paper proposes a new product development model based on AI methods, uses machine learning methods to establish a prediction model of components, processes and properties, and then uses generative models to explore potential compositions and process parameter spaces, replacing the traditional "trial and error method" to accelerate the efficiency of new product development and development. Using the IF steel R&D process as an example, the proposed "digital trial and error method" has greatly reduced R&D costs and time.