According to the complex forming mechanisms, high forming accuracy, and multi-parameters in the bending process of metal tube parts used in rockets and aircraft engines, the parametric FE model for dynamics analysis, forming quality forecast model based on ANN, and process parameters optimization model based on GA were established. A NC tube bending quality forecast and process parameters optimization platform was designed based on these models. Data inconsistency has been solved by data interfaces and process technologies in this platform. By developing corresponding algorithm the special quality indexes extracting function of tube bending was realized. The platform has been used to study forming laws and optimize process parameters of some typical tubes. Management mechanism of empirical data and knowledge was set up. A practical application example was used to show the effectiveness of this platform.